1.1 From Silicon Valley to Stanwick

Governments, universities and land developers take the term ‘megasite’ to mean an enhanced zone with strong infrastructure, shovel-ready for the creation of a new business cluster. The first such megasite, at Stanford University, grew into Silicon Valley. Meanwhile, in prehistoric Europe – a continent with a late development of urban centres in the 1st millennium BC – larger-than-usual sites emerged and disappeared without causing much long-term cultural impact but also without prompting explanation. It is only in the last two decades that megasites have been recognised as an interesting phenomenon worth attention (e.g., in Iberia: Reference Martínez-Sevilla and Lozano RodríguezMartínez-Sevilla et al. 2020). But even when megasites have stimulated research projects, remarkably little attention has been paid to why the sites were so large (e.g., Corneşti: Reference Heeb, Szentmiklosi, Bălărie, Heeb, Szentmiklosi, Krause and WemhoffHeeb et al. 2017). Moreover, and with the exception of Roland Fletcher’s research (Reference Fletcher1995; see Section 1.1), the tendency in megasite studies for investigations in terms of their own local cultural context has precluded comparative approaches.

In this Element, we use a relational, comparative approach to identify not only what made megasites but also what made megasites so special and so large. We address three themes in this Element – megasites in general, the phenomenon of congregation and five examples of special megasites. Congregation consisted of formal meetings at various scales according to punctuated temporality – people congregated at the congregation place and then left. This occurred not only at megasites but also at smaller sites, such as Rondels (see Section 3) but the scale of congregation events at megasites attracted particularly large numbers of the strangers and kinsfolk – those prehistoric forbears of C. P. Snow’s Cambridge fellows and citizens in his Strangers and Brothers series, dealing with life and times in the colleges and the town and examining issues of personal power and integrity. The size of the Element limits our study to a consideration of one class of megasites (the Trypillia megasites) plus one megasite per period (Neolithic, Copper, Bronze and Iron Ages), with each megasite contextualised through a brief socio-cultural introduction (Fig. 1). By the end of this Element, we hope that readers will have come to understand why megasites in European prehistory mattered and how they can help us to reach a deeper understanding of prehistoric societies in Europe.

Figure 1 Map of key sites discussed in book: 1 – Bil’sk; 2 – Nebelivka; 3 – Corneşti; 4 – Csanadpalota; 5 – Munar; 6 – Sântana; 7 – Idjoš – Gradište; 8 – Foggia Plain sites (Passo di Corvo, Motta di Lupo, Masseria Fragella & Posta d’Innanzi; 9 – Marroquíes – Bajos; 10 – Valencina de la Concepción; 11 – Alcalar; 12 – Porto Torrão; 13 – Perdigões; 14 – Villeneuve – Tolosane & St. Michel-du-Touche; 15 – Corent; 16 – Bourges; 17 – Mont Lassois; 18 – Urmitz; 19 – Wiesbaden – Schierstein; 20 – Heuneburg; 21 – Heidengraben; 22 – Kelheim; 23 – Alsónyék; 24 – Stanwick

(source: authors, re-drawn by Lauren Woodard)

Anyone reflecting on a research strategy for a 450 ha site will immediately realise that these sites are challenging to investigate – that we shall always rely on exploration of only a tiny fraction of their surface area. In this sense, writing this account would have been impossible without the huge advances in geophysical investigations in the last two decades, which have provided crucial detail for megasite internal layouts. But, however indispensable, geophysical plots do not provide chronological detail and it is only with targeted excavation and Accelerator Mass Spectometry (AMS) dating through projects such as ‘The Times of Their Lives’ (or ‘ToTL’: Reference WhittleWhittle 2018) that we can develop improved interpretation. These methodological issues remain as yet unresolved for several of the megasites we consider here. However, the point made in the study of Trypillia megasites – that the methodological advances would be compromised without parallel advances in theoretical understanding (Reference GaydarskaGaydarska 2020) – applies just as strongly to other megasites. But what do we mean by a prehistoric megasite?

1.2 Refining the theoretical framework for megasites

Our central proposition for investigation is that megasites constituted particularly large, though diverse, kinds of congregation places. Before we lay out the relational approach to megasites, it is important to define an appropriate terminology, for many terms are used, sometimes interchangeably, sometimes without definition. The five related terms regularly used in discussions of large settlements – ‘nucleation’, ‘aggregation’, ‘agglomeration’, ‘assembly’ and ‘seasonal gathering’ – do not adequately cover the sense of congregations. ‘Nucleation’ is contrasted with ‘dispersion’, often without a temporal dimension, while ‘agglomeration’ refers to the consolidation of a large, usually permanent population size. But not all agglomeration sites were necessarily congregation places if they lacked the motivation or the facilities for major meetings. The more neutral term ‘aggregation’ is used as much for a collection of structures, physical or social, as for large groups of people. The term ‘assembly’ prioritises the meeting aspect of a site without necessarily invoking an appropriately large scale. The term closest in spirit to congregation is ‘seasonal gathering’, in the sense of Wengrow and Graeber’s (OA/2015; cf. OA/Wengrow 2019) demonstration that seasonal hunter-gatherer group size alternated between bands and large networks (viz., fusion–fission strategies), although, even with this term, the scale of meeting is rarely specified. The notion that sizeable ‘seasonal gatherings’ were normal for human populations in the Holocene, as at Göbekli tepe (OA /Schmidt 2006) or Lepenski Vir (OA/Srejović 1972), provides the key long-term context for congregation places.

In other words, not all megasites were congregation places, if they lacked meeting spaces, and vice versa, if they were too small. Moreover, while small sites often shared similar practices with large sites, the key difference was the intensity of the practices at the larger sites. Megasites can be considered as congregational places if they embodied combinations of great size and scale as well as intensity of practices in a relational sense. It is important to emphasise that absolute size can be misleading. For example, the ditches at the Southern French Chassey sites of Villeneuve-Tolosane enclosed ‘only 25 ha’ but this size exceeded the size of small 0.07 ha Chassey settlements by a factor of 400 (OA/Phillips 1982)! The successive phases of the Iron Age site of Stanwick, North Yorkshire (UK) encapsulated the change from a congregation place to a megasite with regional congregations, with the earlier phases lacking the intensity of depositional practices so clearly attested in the final phase (OA/Haselgrove 2016).

We can extend to much of the farming period in Europe the insight of Reference Danielisová and Fernández-GötzDanielisová & Fernández-Götz (2015) that the later prehistoric and early historic world was still principally the world of the common farmer. Reference Haselgrove, Rebay-Salisbury and WellsHaselgrove et al. (2018) concur, proposing that the vast majority of settlements were small farmsteads with populations of fewer than 100 people. For Adler et al. (OA/1996: 403), local communities were aggregations of people, shared risks, inter-dependencies and identities, constituting the highest level of decision-making above the residential kinship level in non-stratified societies. The political implications of this insight is that the basic settlement units for most of prehistoric Europe were relatively autonomous segmentary societies, where the identities of equivalent settlement segments depended on opposition to such segments but whose social reproduction required their existence (OA/Curras & Sastre 2020), and which came together seasonally for larger-scale interaction. There was an evident political tension between decisions more related to the community level and those taken at the congregational scale. Any shift towards specialised site-based practices, such as regional congregations, at megasites would at once have weakened the independence of segmentary communities while strengthening the regional significance of megasites.

One such practice in which differences between segmentary communities and central places was played out was public architecture. A feature of each megasite was the scale and significance of its public architecture, which would have exceeded the scope of any village or farmstead constructions. However, as L. K. Harrison & Bilgen (OA/2019) remind us, architecture could be mobilised as an instrument of hierarchy or become an integrative medium of community building, insofar as the built environment represented the ‘imagined ideologies’ of the residents and visitors. We should recall that people began to self-identify as an integrated entity before – not after – major constructions, betokening a prior change in community ideology (OA/Ryan 2019: 341–2). Reference Munro, Hetherington and MunroMunro’s (1997) inversion of the division of labour to become the labour of division is pertinent to the manipulation of architectural difference: architecture created spatial divisions in the light of the principle that ‘the more divisive a society, the more space is divided’ (OA/Fernández-Götz & Krausse 2016: 15). However, the converse is true of the large open space in many congregation places, where the absence of divisions created a welcoming reception to all participants.

There was also significant variability between megasites and smaller communities in the way that such constructions constituted ‘the choreography of power’ (OA/Osanna 2016: 275) – an idea related to Sharples’s (OA/2007: 179–80) notion of ‘labour as potlatch’. It will also be important to investigate the extent to which megasite ideological order was driven by a rhetoric of stability and permanence (OA/Riva 2020), as found in site planning, continuities in traditional architectural use of space or the monumental scale of ditches and ramparts, or whether, by contrast, there were important tensions between the maintenance of this rhetoric and its disruption through socio-economic changes.

Another long-term tension found in most, if not all, of the megasites featured here concerns the relations between the domestic and mortuary domains (OA/Chapman 1991) and, in particular, bodily mobility. In the context of well-established settlements, there is an assumption that most or all of the local deceased were buried nearby in specially constituted, permanent cemeteries. However, mobile forager sites and dispersed farming homesteads had a tethered relationship to their central cemeteries, with bodies regularly moved from the (sometimes seasonal) settlement to their permanent, ancestral place (e.g., megalithic tombs in much of western Europe). There is therefore a widespread precedent in time and place for the transport of the complete bodies of the deceased for burial across distances of 5–10 km to cemeteries and megalithic tombs. Another term for these places is a mortuary congregation. We should therefore expect to identify congregational site in the mortuary as well as in the domestic domain.

It is precisely the mortuary domain to which Munro’s notion of the labour of division is pertinent in both the spatial and ontological senses. The centrality of the categorisation of the human body in these practices implies that the more varied the mortuary practices, the more different the ontologies of the persons buried. One way in which mortuary variability was played out was through the tension between the dominant mortuary practice and variations on, or oppositions to, the main practice. If we can recognise in the megasites a trend through time towards greater variability of mortuary practice, did this correlate not only with social differentiation but also with individualisation – with mortuary personalisation?

In summary, megasites as congregational places can be considered as special combinations of relational size, scale and intensity of practices. The sites that we shall examine here are examples of Fletcher’s intermediate group of low-density settlements of between 1 and 100 km² (Reference FletcherFletcher 2019: 17). There is a long-term Holocene contrast between smaller residential sites and larger (sometimes much larger) seasonal congregational sites. A major issue for consideration is the extent to which the community’s loss of political autonomy (autarky) was offset through the scalar advantages pertaining to not just large but massive congregational centres. In the next section, we consider the relational approach to megasites as congregational centres.

1.3 The Relational Approach to Megasites

It will have become clear by now that the relational approach to megasites forms the foundation of the Element. One of the authors (B. Gaydarska) has developed this approach in her study of the low-density urbanism characteristic of the Trypillia megasites (OA/Gaydarska 2016; Reference GaydarskaGaydarska 2020, chapter 6) in an attempt to circumvent the inflexibilities of those urban studies dominated by high-density cities. Here, although the urban theme is not pursued further, the touchstone of the relational approach is what people experienced in ‘normal’, small settlements and what they experienced in megasites. It is a fundamental categorical mistake to assume that megasites were settlement sites writ very large; this is tantamount to assuming that aircraft carriers were simply very large versions of yachts.

One of the most complex issues for urban sites and megasites alike remains the currencies used for comparison. Standard units of analysis, such as Childe’s criterion of writing (OA/Childe 1950), can readily become ossified and essentialised, while others, such as ideology or social power (OA/Cowgill 2004: 543), are too general for differentiating urban from non-urban. The parameters chosen in this study are as follows: scale, temporality, deposition/monumentality, formal open spaces, performance and congregational catchment. We claim that these six parameters provide a common frame of reference and an appropriate terminology for all of the megasites discussed here in detail. While accepting that there are other parameters that could have been considered, we suggest that there are good research reasons for the inclusion of these parameters.

1.4 Structure of the Element

The structure of this Element is straightforward. Sandwiched between an introductory section and a discussion and summary of the principal findings of the research (Section 7) are five sections in which, with one exception, a general introduction to the period in question precedes a more detailed analysis of a single, representative megasite (Sections 2–6). Six detailed analyses – the mortuary and land use modelling and the construction of ‘houses of the dead’ at Alsónyék, the chronological analyses of Valencina de la Concepción and the labour studies at Corneşti and Bil’sk – are placed in the Online Appendix, together with a more detailed account of the European Iron Age and much of the further reading (references beginning OA/, as in OA/Anderson 1991).

In Section 2, we begin with a recapitulation of a whole class of megasites – the 4th millennium cal BC Trypillia megasites. Since we have already published a detailed study of the Nebelivka megasite (Reference GaydarskaGaydarska 2020), we offer a comparison of the three largest Trypillia megasites – Taljanki, Nebelivka and Majdanetske.

In Section 3, we go back in time to the 5th millennium cal BC – to the Late Neolithic and Chalcolithic of the Balkans and the Carpathian Basin and the Late Neolithic of Central Europe. Despite a range of potential megasites with evidence for settlement congregations, we have chosen the Lengyel site of Alsónyék, in south-western Hungary, as an example of a mortuary congregation place that resulted in the largest mortuary complex known in Neolithic Europe.

Our choice of megasite for discussion in Section 4 is Valencina de la Concepción, near Sevilla, in Southern Spain – a site whose precise chronological sequence (3200–2300 cal BC) resulted from its inclusion in the ‘ToTL’ Project. The size of the complex and the diversity of its finds at Valencina made this an obvious choice when considered in comparison with the smaller megasites and other enclosures of Neolithic and Copper Age Southern Europe, including Italy, southern France and southern Iberia.

In the utterly different world of the European Bronze Age, dominated by small hillforts and peopled by dispersed homestead groups, it is rare to find any megasite at all. Once a landscape marked by mega-forts had been identified in the Carpathian Basin, it was clear that one site stood out from all others in the 2nd millennium cal BC for inclusion in Section 5 – the huge enclosed site of Corneşti.

The final choice of a 1st millennium cal BC, Iron Age megasite for treatment in Section 6 was also surprisingly straightforward, even though the size and complexity of Late Iron Age oppida in western and central Europe were well known. Our selection returns us to the Ukraine, where the largest hillfort in Europe was founded at Bil’sk (aka ‘Bel’sk’) around the time of the earliest Greek colonies on the North Pontic shore. This choice was made particularly easy because of the almost total neglect of the site (a single, three-line mention) in the Oxford Handbook of the Europen Iron Age (Reference Haselgrove, Rebay-Salisbury and WellsHaselgrove et al. 2018).

While other localities could undoubtedly have been proposed for inclusion, we are content that each of these five examples constitutes a place that conveys important aspects of the period in question. In other words, these five megasites are not ‘representative’ of their millennium in a statistical sense but in the cultural sense of a central place whose congregational practices ‘represent’ the wider regional society in which it is embedded. It is for this reason that we remain confident that this choice of five megasites not only provides a key to understanding the creation and maintenance of the megasites themselves but also helps us to move forward in a broader understanding of European prehistory.

2.1 Introduction

The Trypillia group was but one part of the wider Cucuteni-Trypillia group, whose immense duration of over two millennia (4800–2700 cal BC) was matched only by its extreme size of cca. 250,000km² (Reference VideikoVideiko 2013; Reference GaydarskaGaydarska 2020) (Fig. 2). Although there were many shared features of Cucuteni and Trypillia material culture in pottery, figurines and houses, there was a divergence in settlement trajectories between the two groups, with Cucuteni sites rarely covering more than a few hectares but with Trypillia sites growing to 10 ha in Phase A, 100 ha in Phase BI and 236 ha in Phase BII (Fig. 3). Sites of 100 ha or more have been dubbed ‘megasites’ (Reference Chapman, Videiko and HaleChapman et al. 2014), the largest of which (the Phase CI Taljanki) reached 320 ha (OA/Rassmann et al. 2014) (Fig. 4). The megasites clustered in the interfluves between the Southern Bug and Dnieper rivers, with outliers to the east and west.

Figure 2 Map of key Trypillia-Cucuteni sites, with sources of copper, salt, manganese and Gumelniţa pottery: 1. Grebeni; 2. Stolniceni; 3. Petreni; 4. Vesely Kut; 5. Apolianka; 6. Dobrovody; 7. Maidanetske; 8. Nebelivka; 9. Taljanki; 10. Haivoron graphite

(source: M. Nebbia and Nebelivka Project)

Figure 3 (upper) Plan of Nebelivka + Quarters; (lower) plan showing size of Grebeni relative to megasites

(source: authors)

Figure 4 (upper) Plan of Taljanki; (lower) plan of Maidanetske

(source: OA/Rassmann et al. 2014, figs. 3.1a and 22a)

Although Trypillia settlements were discovered at the turn of the 19th century (OA/Khvoika 1901), megasites were not identified until the ‘first methodological revolution’ of the 1960s–1970s, when aerial reconnaissance revealed the complex plans of hundreds of burnt houses, which, on excavation, were found associated with Trypillia pottery (Reference VideikoVideiko 2013). After forty years of mostly house excavation, the ‘second methodological revolution’ was initiated in the late 2000s, with international teams using remote sensing, AMS dating and a battery of palaeo-environmental techniques to access new information (Reference Chapman, Videiko and HaleChapman et al. 2014; Reference Müller, Rassmann and VideikoMüller et al. 2016). The primary focus of the more precise geophysical investigations remained megasite plans (Figs. 3–4), still dominated by burnt houses but also including new features such as unburnt houses, communal buildings termed ‘Assembly Houses’, pits, industrial structures and pathways. Supra-household configurations showed two levels of site structure – Neighbourhoods comprising three to twenty-seven houses and Quarters comprising five to eighteen Neighbourhoods. The new plans also showed marked architectural variability, suggesting a bottom-up approach to planning that had hitherto been overlooked.

However, these methodological breakthroughs were necessary but insufficient advances to gain a deeper understanding of megasite communities, not least to answer the question of urban status. Gaydarska (OA/2016; 2020, chapter 6.3) developed a relational approach to megasites through the structured comparison of the Nebelivka megasite with the small Trypillia settlement of Grebeni – an approach that provides the framework for this Element (see Section 1) (Fig. 2). In a later study, a detailed comparison of the six parameters used in this Element was made for the three Trypillia megasites of Taljanki, Nebelivka and Maidanetske (Reference Gaydarska, Chapman, Diachenko, Harper, Rassamakin and Sobkowiak-TabakaGaydarska & Chapman 2021) (Figs. 3–4). The remainder of this section summarises the main points made in this recent research on the congregational significance of the Trypillia megasites.

2.1.1 Scale

The size of the majority of Trypillia sites in each Phase was 10 ha or less (Reference Nebbia and GaydarskaNebbia 2020) – a reminder that small community size and limited site size were central to the Trypillia habitus. The appearance of megasites was utterly unexpected for their communities, with the increase in scale proving transformative on all levels. This was as true for the number of houses constructed and burnt each year as it was for the building and moving of Assembly Houses from one Neighbourhood to another. It was also true for the construction of the largest Assembly House at Nebelivka – the so-called Mega-structure (Fig. 5) – with its great size matched only by the massive assemblage of mostly fragmented vessels deposited before its final burning. And it was also true for the large number of Quarters and Neighbourhoods planned for Nebelivka – each Quarter larger than a ‘normal’ small Trypillia settlement site and each Neighbourhood the living space for the visitors from a single small settlement.

Figure 5 (upper) Reconstruction of Nebelivka mega-structure (source: C. Unwin and Nebelivka Project); (lower) House-burning, Nebelivka

(source: Nebelivka Project Archive)

2.1.3 Deposition and Monumentality

If the core of Trypillia archaeology is the selective deposition and fragmentation of objects, deposition lies at the heart of punctuated temporality. The variability of megasite deposition concerns the context of the event – whether private (in a house) or public (in a large pit or the site’s perimeter ditch) – as much as its timing (inside a house before burning or placed on the top of the burnt remains) and the scale of the practice, which ranged from placing fragments of ten vessels in a pit to the filling of the largest Assembly House at Nebelivka (Fig. 5) with parts of 332 vessels before the two-stage burning of the structure. Participation in the deposition event revealed contrasts in relations between persons and households, with deposited objects presencing other places, other persons and other times.

The experience of monumentality varied with perspective, whether witnessing the sounds, smells and sight of a megasite in its entirety from a neighbouring hill or the local appreciation of a ceremony such as a procession towards an Assembly House or a two-storey dwelling house in flames. The cumulative growth of minor mortuary landmarks such as the memory mounds (Fig. 6) transformed Nebelivka from a place of the living into a meeting place for the living and the ancestors.

Figure 6 Memory mounds and inner streets, Nebelivka: clockwise from upper left: memory mound, TP 24/4; memory mound, TP 22/4; square and blocking streets, Quarter G; parallel inner radial streets with blocking street, Quarter B; converging inner radial streets, Quarter L

(source: Nebelivka Project Archive)

2.1.5 Performance

Performance was an essential part of the Trypillia habitus, closely related as it was to the timemarks of the social calendar. What differentiated megasite performance from those at smaller sites was the significance of major ceremonies. The temporality of performance included three stages – preparations, the performance itself and the incorporation into social memory. These timemarks took four principal forms: house-burning, feasting, other kinds of deposition and processions. As the Nebelivka Project discovered during their experimental house-burning (OA/Gaydarska et al. 2019), the stunning visual effects of house conflagrations (Fig. 5) were amongst the most memorable events of the calendar, not diminished by their regularity and notable for the unique scale of Assembly House burning. All four forms of feasting as delineated in Reference KassabaumKassabaum’s (2019) feasting typology could be identified at Nebelivka through the varied form and scale of animal bone deposition. Object deposition formed the centrepiece of hundreds of other depositional performances in all the main megasite contexts. The well-established principles of megasite planning, including the space between the two principal house circuits as well as the parallel inner radial streets leading into the inner open area (Fig. 6), conformed closely to the performance of processions into and around the megasite (Fig. 7).

Figure 7 Singing, dancing and processions: (a) dancing pot Chirileni III (Reference Gaydarska, Chapman, Diachenko, Harper, Rassamakin and Sobkowiak-TabakaGaydarska & Chapman 2021, fig. 2.6a, based upon OA/Monah 2016, fig. 268/1); (b) Kolomiishchina singing figurines (Reference Gaydarska, Chapman, Diachenko, Harper, Rassamakin and Sobkowiak-TabakaGaydarska & Chapman 2021, fig. 2.6b, based upon Ciuk 2008, p. 227); (c) singing figurine, Maidanetske (Reference Gaydarska, Chapman, Diachenko, Harper, Rassamakin and Sobkowiak-TabakaGaydarska & Chapman 2021, fig. 2.6c, based upon OA/Ciuk 2008, p. 215); (d) dancing scene, Brânzeni (Reference Gaydarska, Chapman, Diachenko, Harper, Rassamakin and Sobkowiak-TabakaGaydarska & Chapman 2021, fig. 2.6d, based upon OA/Monah 2016, fig. 249/3); (e) Nebelivka procession

(source: C. Unwin & Nebelivka Project)

2.1.6 Congregational Catchment

The interpretation of a congregational catchment relates to the people visiting the megasite and the objects that arrived there. In view of the absence of human burials at megasites, Reference Nebbia and GaydarskaMarco Nebbia (2020) has developed formal settlement modelling to demonstrate that the majority of megasite interactions took place in up to a 100 km radius. In the sense that the movement of people living more than 15–20 km from a megasite meant overnight stays at smaller settlements, the congregation process led to a consolidation of social relations within the catchment.

Early Trypillia communities were connected to the jadeite network linking the French Alps and Eastern Europe, with two axes known from dated contexts: Slobidka – Zahidna (Phase AIII) and Berezivka (Phase BI) (OA/Pétrequin et al. 2017). In the later megasite phase, no more jadeite is known but small quantities of copper, gold, salt, chipped stone and manganese pigment (Fig. 8) came from beyond the 100 km catchment in two stages. First, the exotics were brought into the catchment, from which local visitors brought them to megasites. Later, with the expansion in the fame of congregations, long-distance specialists (sensu OA/Helms 1988) had the opportunity to bring exotics directly to the megasites. Such exotics were one of the many attractions for visitors to megasites; no wonder visitors took them home instead of depositing them at the megasite.

Figure 8 Exotics at megasites: (a) copper axe, Maidanetske (source: OA/Ryndina 1998, Ris. 66/6, re-drawn by Lauren Woodard); (b) copper awl, Taljanki (source: OA/Ryndina 1998, Ris. 66/12, re-drawn by Lauren Woodard); (c) Nebelivka graphite-painted dish with internally thickened rim; (d) Nebelivka gold ornament (7 x 3 mm); (e) Prut – Dniester flint, TP 19/2; (f) Nebelivka painted vessel (manganese) TP 1/3

(source for (c)–(f): Nebelivka Project Archive)

2.2 Conclusion

Trypillia megasites can be considered as an exemplar of congregation sites for European prehistory, with many of the six variables discussed here appropriate to an analysis of later potential megasites. One of the outstanding features of Trypillia megasites was the combination of a major dwelling zone with a massive open congregation area. This combination shows that the supposed mutual exclusivity between dwelling and ritual/deposition is a false dichotomy. There is also no contradiction between a congregation centre and a site of urban status, since the former influenced all the performances of the latter.

3.1 Introduction

The Balkans and Central Europe resemble the Cucuteni-Trypillia distribution in the domination of their Neolithic archaeology by settlement remains. The mortuary domain was relatively minor in most areas and, until the Fall of the Berlin Wall opened up the floodgates of remote sensing in these regions, there were relatively few known public monuments, except for the Late Neolithic Hohensiedlungen (hill-top sites) and Rondels (circular enclosures) of Central Europe (OA/Petrasch 2015). By the 2020s, our understanding of the Neolithic landscapes of Hungary and South East Europe has been transformed by the discovery of large numbers of enclosures, some built in combination with Balkan site types such as tells. However, even using the relational approach to site size and scale, we can identify relatively few megasites in regions where nucleated settlements are common (Table 1). Why is it that there were so few megasites in such settlement-rich landscapes?

A well-known trait of Neolithic lifeways is a reduction in mobility, whose reduction in direct access to ‘resources’ was mitigated by an increase in exchange practices between more settled places. This created a new form of the Palaeolithic problem of the ‘absence’ of the vast majority of people known by a community, because they were living elsewhere. One common Neolithic response in increasingly permanent communities was the creation of seasonal meeting places.

In the Earlier Neolithic (6300–5300 BC), we can distinguish between tell landscapes in North Macedonia and Bulgaria and landscapes dominated by ‘flat’ sites (Romania, Central Balkans, Hungary) (OA/Reference ChapmanChapman 2020a) (Fig. 9). The Bulgarian tells are a classic example of what we have discussed as segmentary societies (see Section 1.2) – mostly small settlements defined by often low mounds, with site populations of 50–75 people. The mating networks required for long-term viability would have needed the participation of 10–12 tells. With high densities of houses on the tell, ‘off-tell’ meeting places were the obvious solution but there is little tradition of excavating away from the tell in these countries. One of the few Earlier Neolithic enclosures in a tell-dominated landscape was at Yabulkovo, in South Bulgaria (OA/Roodenberg et al. 2014) (here Fig. 10a) – a triple-ditched enclosure with deposition in pits in the central area.

Figure 9 Map of sites mentioned in Section 3

(source: authors, re-drawn by Lauren Woodard)

Figure 10 Early Neolithic Yabulkovo: (a) plan (source: Leshtakov 2014, fig. 77, re-drawn by Lauren Woodard) and ditch B1

(source: Petrova 2014, fig. 3.7); (b) geophysical plan of Drenovac (K. Rassmann)

In the non-tell world of the Starčevo – Körös – Criş groups, settlement varied between one-house sites, hamlets and villages (OA/Chapman 2008). Village sites such as Alsónyék, in Hungary, Leţ, in Romania or Galovo in Croatia, may have included meeting places within their settlement but these are not obvious. This leaves enclosed sites such as Cârcea, in South West Romania, as a potential congregation place.

In the Later Neolithic (5300–4600 BC), amidst rising regional settlement densities and the expansion of tell lifeways, there was an increased diversity of site types and sizes, with larger sites showing a greater variety of combinations of site elements such as the mound, the horizontal settlement, the Rondel and the enclosure (Reference ChapmanChapman 2020, chapter 5). But the basic settlement unit reproducing segmentary societies was the small tell or flat site of <2 ha area with a population of <100 people. This was true of most phases of the Karanovo sequence in South Bulgaria, with the exception of phase IV, when most people lived on flat sites and created meeting places characterised by extensive deposition of ordinary day-to-day objects in pits.

In places where increased sedentism had led to greater population nucleation (Table 1), the concentration of so many people perhaps meant that off-site assembly places were not necessary. However, the geophysical plots of sites such as Drenovac (Fig. 10b), Pločnik and Belovode show few signs of a meeting place within the settlement. It should be noted that the largest Vinča sites were not as big as proposed earlier (OA/Chapman 1981). On the basis of new fieldwork, the 100 ha size of Potporanj refers to an area of 1 sq km where settlement was in fact highly discontinuous; equally, the 80 ha size of Selevac can be revised downwards to 53 ha and the 65 ha of Turdaş to 23 ha. By contrast, Csőszhalom included two special assembly foci – a tell and an enclosure – within a large horizontal settlement (Fig. 11). Elsewhere, combinations of a tell within, or adjacent to, an enclosure at sites such as Uivar, Szeghalom and Bordjoš illustrates how meeting places could be incorporated into site plans.

Figure 11 Plan of the Csőszhalom complex

(source: OA/Raczky 2018, fig. 3.1)

In Transdanubia and Croatia, a far higher density of enclosed sites was found in the Sopot and Lengyel groups than in the Central Balkans. Two sets of enclosures are known – the Djakovo (OA/Šošić-Klindžić et al. 2019) and Szemely (Reference Bertók and GátiBertók & Gáti 2014) groups, each with remarkably different enclosure morphologies. In the former, the site of Tomašanci comprises five separate small Sopot enclosures, while the largest site of Gorjani – Kremenjača – revealed multiple dwelling features within three concentric ditches. By contrast, most of the enclosures in the Szemely group were Rondels of widely differing complexity (Fig. 13). The presence of an enclosure or an empty space at the centre of several Sopot settlements suggests regular on-site assembly places. An unusual combination concerns the creation of a Rondel in the middle of the Lengyel settlement site of Zengővárkóny, with its clusters of houses and burials. This site forms a major contrast to Alsónyék, which lacks a Rondel (see Section 3.2). A settlement that does not obviously fit these Central Balkan-Carpathian patterns is Iclod, in Transylvania, with its so far unique combination of enclosure, settlement remains and mortuary zone.

Figure 12 (a) Plan of Urmitz; (b) bastions at gates, Urmitz; (c) Hut 1, Urmitz; (d) Hut 6, Urmitz; (e) plan of Wiesbaden – Schierstein

(source: OA/Boelicke, U. 1976/7, figs.1.1, 2.2–2.3, 2.6 and 6.7/4: re-drawn by Lauren Woodard)

Figure 13 (a) Lengyel sites in the Carpathian Basin (source: OA/Scharl 2016, fig. 2.4), with enclosed sites in Co. Baranya (source: Reference Bertók and GátiBertók & Gàti 2014, II.90) (combined and re-drawn by Lauren Woodard); (b) Soil environment of Alsónyék and Mórágy

(source: Depaermentier et al. 2020, fig. 2.1b, re-drawn by Lauren Woodard)

The emergence of cemeteries in the late 6th millennium cal BC created an alternative domain for the negotiation of ancestral links to that of the well-developed domestic domain (Reference ChapmanChapman 2020, chapter 6). Cemeteries were rare until c. 4800 cal BC, even though large complexes such as Cernica (Romania), with c. 378 burials, were known from early on. The only region with consistent cemetery usage from 5300 cal BC onwards was the Western Black Sea coast (Hamangia group). This formed the basis for exceptional cemeteries such as Varna, with the world’s first concentration of gold objects (OA/Ivanov & Avramova 2000), and Durankulak – until Alsónyék, the largest aggregation of burials in South East Europe (OA/Todorova 2002). Alongside cemeteries of 30–100 burials, there were rare instances of mortuary congregations, such as the sample of over 10,000 bones (human and animal) deposited in four pits at Alba Iulia – Lumea Noua, representing over 100 human individuals whose bones were brought from other sites for central burial (OA/Gligor 2009). Elsewhere, in the Lengyel group, clusters of burials lay close to sets of houses in a pattern fundamentally different from the small- to medium-size cemeteries of the East Balkans and reaching 368 burials at sites such as Zengővárkóny (OA/Dombay 1960).

The general conclusions are that, for the Earlier Neolithic, there were few examples of settlements with obvious meeting places and even fewer special and separate meeting places. For the Later Neolithic, meeting places were often defined within the overall boundaries of large sites, even if the site was dominated by everyday dwelling. The exceptions were Karanovo phase IV pit sites and sites in any region with a high proportion of small dispersed settlements.

The largest sites in the Balkan-Carpathian regions were between twenty and forty times as large as the usual small (<2 ha) dispersed segmented settlement. There was undoubtedly an up-scaling factor in site practices at these larger settlements, which nonetheless seemed to fall below the threshold of megasites. Many of these long-lasting Balkan settlements formed key nodes in stable exchange and marriage networks, perhaps displacing the need for even larger sites. If the Trypillia megasites emerged in the context of agro-pastoral frontier conditions during the eastwards Trypillia expansion (see Section 2), we should perhaps look for megasites in the initial Neolithic expansion into South East Europe. However, the restriction of the site sizes during the earliest farming expansion to a tenth smaller than those of the Early Trypillia sites meant a far lower starting point for site growth, leading to the absence of megasites in the Balkan Early Neolithic. The longer-lived cemeteries, with their hundreds of bodies, offered an alternative ancestral focus to that of the long-lived tells and flat sites.

Further north and west, settlements of the Early Neolithic group termed the Linearbandkeramik (or LBK) spread as far as the North Sea and into the Paris Basin (Reference Bickle and WhittleBickle & Whittle 2013). As with the first farmers of the Balkans, LBK sites ranged from single-house sites and hamlets to villages. There is currently little evidence for special meeting places in this group, with the exception of the mortuary congregation at the Late LBK Herxheim enclosure (OA/Chapman et al., in press). However, the increased intensity of enclosure, which had already started in the LBK, was a major characteristic of the Late Neolithic, ranging from small Rondels to the two megasite enclosures of Urmitz and Wiesbaden – Schierstein, in the Rhein valley – both c. 100–120 ha in size. Although sharing many constructional and depositional details with Michelsberg and other North European enclosures, these two sites appear to form a class of their own (OA/Petrasch 2015).

Both sites share a semi-circular enclosure form, with ditches 2.5 km long at Urmitz (Fig. 12) and 1.5 km long at Wiesbaden (Fig. 12). Some of the multiple causeways spaced roughly every 100 m at Urmitz had complex entrance-works (OA/Boelicke 1976/7) (here Fig. 12 inset). While there was frequent deposition of pottery in the ditches at both sites, with burials in the ditches at Urmitz, Wiesbaden showed a wider range of deposition in the interior than Urmitz, with some evidence for on-site antler-working and flint-knapping. Small rectangular houses were found in the interior at Urmitz alongside over 75 pits. No explanation has been provided for the extreme size of these enclosures, although Gronenborn et al. (OA/2020) have proposed links between Michelsberg enclosures, local hierarchies and the exchange of salt and jadeite axes. While Gronenborn (p.c., February 2021) has confirmed the rarity of coeval sites in the Urmitz catchment, an explanation of congregation sites should be considered for Urmitz and Wiesbaden – Schierstein.

3.2 Alsónyék: A Mortuary Congregation Site?

Alsónyék-Bátaszék is located in Transdanubia in modern-day Hungary and associated with several Neolithic cultural groups (Starčevo, LBK, Sopot and Lengyel). The site was excavated in advance of the M6 motorway in 2006–9 and produced 15,000 features (Reference Osztás, Bánffy and Zalai-GaálOsztás et al. 2016a) (Fig. 14). It is an extraordinary site on five grounds. It is the longest-lived flat site in the European Neolithic, with over a millennium of uninterrupted occupation (5350 cal BC to 4300 cal BC) that is comparable only to tells. Its estimated size of c. 50 ha is the largest known Lengyel occupation. Its 2,300+ Lengyel graves constitute an unparalleled burial phenomenon. It was extensively excavated over an area of 25 ha, showing huge numbers of houses, burials, pits and pit-complexes and over 500,000 sherds. Finally, the site benefits from exemplary applications of best practice in archaeological science (e.g. Reference Bayliss, Beavan and HamiltonBayliss et. al 2016; Reference Depaermentier, Kempf, Bánffy and AltDepaermentier et al. 2020a, Reference Depaermentier, Osztás and Bánffy2020b). Almost every aspect of this amazing site has important implications for our understanding of cultural developments in Western Hungary. There is an ongoing publication programme of the resultant massive corpus of data and analysis, with some crucial results already published (Reference Bánffy, Oross and OsztásBánffy et al. 2016). We concentrate here only on the published data on the Lengyel occupation phase, with its 122 houses and over 2,300 burials, representing 90 per cent of all Alsónyék graves.

Figure 14 Plan of Alsónyék complex

(source: Reference Osztás, Bánffy and Zalai-GaálOsztás et al. 2016a, fig. 2.6 and OA/Rassmann 2020, fig. 4.8, with additions by Lauren Woodard)

The unprecedented number of burials led to the hypothesis of an equally vast ‘coalescent community’ living at Alsónyék (Reference Bánffy, Oross and OsztásBánffy et al. 2016). Details about our insights on this model and the pathway to the alternative suggestion of mortuary congregation are provided online (Online Appendix I). Our starting point is that the relationship between burial and settlement is not straightforward, as burial preceded the start of dwelling activities and ceased before the end of the Lengyel settlement. This suggests that it is the huge concentration of burials that defines this site. In short, instead of the emergence of a coalescent community of more than 2,000 people over fifty years, we see a process of nucleation of half that number of people – still unprecedented in the Lengyel context – only part of whom were buried in Alsónyék alongside people who were brought to Alsónyék as their final resting place. It is possible that such a burial tradition started in the Sopot period, which may have overlapped with the first Lengyel burials; what is certain is that the earliest Lengyel burials were made without accompanying dwelling practices (Reference Osztás, Zalai-Gaál and BánffyOsztás et al. 2016b: 223). Thus the main attraction for the initial dwelling and subsequent nucleation was the increasing importance of Alsónyék as an ancestral mortuary space whose cumulative burials emphasised place-value in the way that repeated dwelling did on tells. Alsónyék’s advantage over other places was to provide a cohesive, rather than a competitive, space for the display, negotiation and reproduction of social identities at both the local and regional levels.

3.2.4 Performance and Open Area

The availability of formal open spaces in the Lengyel complex seems to have been limited. Performance on the other hand was a frequent experience, involving grave digging, body manipulations (carrying, placing in position, adorning), making sacrifices, food preparations, grave good deposition, chanting, crying, singing or other mourning and celebratory practices. The construction of the unique houses of the dead (Fig. 15) started early in the sequence and continued late. They inevitably concentrated the experience at the time of burial but acted as a constant visual focus of commemorating rituals and ceremonies. The relationship of these houses to the identity of the deceased and their wider role as a structuring element in the ceremonialism of an ever-expanding burial ground cannot yet be assessed. We do not yet know the complete distribution of houses of the dead by grave group but the dated examples show visually impressive clustering in a few grave groups (e.g., Groups 13 and 56), with single houses in other groups (e.g., Group 14 and 57). However, their restriction to Alsónyék underlines the site’s importance by combining the symbolism of both a landmark and a timemark in the formality of the burial rituals. The prominence of large bovids, both wild and domestic (Reference Nyerges and BillerNyerges & Biller 2015), in the faunal sample emphasises the importance of communal feasting, whether at the grave-side or in domestic practices.

Figure 15 (b)–(d) Plan and section of Grave 10B/6537, Alsónyék (source: Reference Zalai-Gaál, Osztás and KöhlerZalai-Gaál et al. 2012, Abb. 8); (a), (e)–(f) reconstruction of its House of the Dead

(source: Stuart Johnston)

Another important aspect of performance concerns movement. The context of movement at Alsónyék was the sheer size of the complex, requiring twenty to thirty minutes to cross the 1.5 km-long site. Those walking from one end to the other would have passed by over fifty grave groups, many with houses of the dead at their core, as well as many clusters of dwelling houses. Moreover, the arrival of the deceased from other settlements was doubtless accompanied by formal processions at least within the site perimeter, which probably culminated at a ceremony outside a house of the dead. These processions cemented the growing desire of ‘outsiders’ to be buried at Alsónyék and enhanced burial performance.

3.3 Conclusions

Rather than a coalescent community of which an excessively high proportion would have required burial, we propose that Alsónyék was primarily a place of mortuary congregation for locals as well as deceased brought from other settlements. This practice began on a small scale, with an enormous expansion c. 4700 cal BC, bringing Alsónyék a regional fame for burial. This striking form of congregation site differed as much from the Trypillia megasite congregations of the living as it did from the megasites of the Central and Western Mediterranean, to which we now turn.

4.1 Introduction

In Section 3, we focussed on German megasites with little evidence for dwelling and a Lengyel megasite dominated by a mortuary domain juxtaposed with intensive dwelling. In this chapter, we move westwards into Southern Europe to examine three regions whose sites show some parallels with the Central European megasites (Table 3). In Southern Italy, ditched enclosures formed the primary settlement form in the Tavoliere plain from the earliest known farming period, with the megasite phenomenon occurring only after several centuries. By contrast, ditched enclosures were rare in the Early Neolithic of Southern France and Southern Iberia, with megasites occurring in later periods characterised by more settlement aggregation, elaborate material culture, striking mortuary remains and the formation of regional polities. The regional chronologies are presented below (Fig. 16).

Figure 16 Regional chronologies, Southern Italy, Southern France and Southern Iberia

(source: authors, re-drawn by Lauren Woodard)

Manuel Fernández-Götz’s characterisation of the Iron Age as: ‘the more divisive the society, the more space is divided’ (OA/Fernández-Götz & Krausse 2016, 15) is relevant here. Space was divided up with visible features such as banks and palisades or less visible cut features such as ditches, often interrupted. Perimeter banks and house walls were rare, while palisades were found on only a few parts of the two large Chasséen enclosures. The vast majority of division was accomplished through digging ditches, which demarcated significant places in the landscapes and structured access to/from interiors. The most important element of a cut feature was the relationship it created with the past – whether earlier cultural features (the ancestral past) or with natural sediments (nature) (OA/Chapman 2000). Digging a pit or a ditch into nature meant establishing an exchange with an undifferentiated past; the taking of ancestral natural materials, with the exchange deferred until later deposition in the cut feature. In the case of Perdigões (Portugal), the performance of excavating one ditch by extracting 14,323m³ of rock was followed by removal of the rock from the circuit rather than the construction of a bank (Reference Valera, In, Gibaja and ChambonValera 2012), leaving a ditch that welcomed many special deposits. The pits inside many of the Iberian enclosure ditches usually contained special deposits, whose elaboration in terms of materials used and sequence of action increased with time. Such exchanges were central to both large and small sites, even in small unenclosed sites with pit deposition.

There was an important temporal tension between permanence and change at the larger sites. While Valera (OA/2006) has emphasised the permanent state of flux at enclosures, with occupants sharing new experiences as design merged with new building, site duration sometimes exceeded a millennium, with continuity of place related to the ritual or cosmological focus. Moreover, the excavation of many site ditches in segments meant different groups mobilising their own construction labour (OA/Díaz-del-Río 2006). Such constructions were dispersed in time as well as space, with a network of social relations of mutual support based upon feasts, gatherings and projects. It is thus not surprising that, the longer the site use, the more probable the polyfocal nature of performance.

In his discussion of Iberian enclosures, Marquez (OA/2003) denies the general separation of sacred and profane. However, the dialectic of this relationship varied through time and place: indeed, the relationship between dwelling and burial became problematic. In the Copper Age at Perdigões, the boundary between the living (‘domestic’ discard) and the dead (bodily discard or burial) became blurred over time through the diversification of mortuary practices, which increasingly became the site’s key performances. It is intriguing that communities used the commingling of sacred and profane to transcend their own spatial divisions sensu Fernández-Götz! Thus we can re-evaluate in two ways Reference Hurtado, Valera and Shaw EvangelistaHurtado’s (2006) striking claim that enclosures acted as a monumentalisation of domestic space. The first is that the profane must be included in many aspects of this vertical up-scaling process. The second is that ditches – even 6 m deep ditches – could have been visible ‘monuments’ only on a local scale of 10s–100s of metres. There is an evident tension between social memory and visibility in ditched enclosures, which was worked out locally in diverse ways. We now turn to the three regional sequences.

4.2 Southern Italy

There is a long tradition of enclosing small spaces of up to 4 ha from the start of the Neolithic settlement on the Tavoliere, which cumulatively formed one of the densest site concentrations in Neolithic Europe (OA/Jones, G. D. B. 1987; OA/Hamilton & Whitehouse 2020). These ditched enclosures were often multi-period sites, with up to eight concentric ditches. Only at a later stage, communities started to excavate long, if irregular, stretches of single ditches to enclose what have been called ‘Annexes’ but which were actually the key part of the new extended sites. In fact, only a handful of large sites, and even fewer small sites (<4 ha), consisted of both outer and inner (settlement) enclosures.

The largest of these is the megasite of Passo di Corvo, whose main enclosure of 38.5 ha and outer enclosure (annexe) of 64 ha produced a total size of 104.5 ha (Reference Seager and ThomasSeager Thomas 2020) (here Fig. 17). G. D. B. Jones (OA/1987, 101) described Passo di Corvo as ‘the major prehistoric settlement of the Tavoliere’. The key difference between the outer and inner enclosures is the density of smaller C-shaped ditches in the latter, accompanied with intensive deposition, with a relative paucity of C-shaped ditches or finds in the former. The outer enclosures were largely empty spaces – effectively equivalent to the central open spaces in Trypillia sites (see Section 2). Previous authors have proposed defence, stock enclosures and fields for the outer enclosure. We propose, instead, that their primary function was for seasonal congregation for people from the local and other communities. The Passo di Corvo outer enclosure would have attracted visitors from a social catchment of the entire Tavoliere, estimated at 5,000 people (OA/Robb 2007), replacing the earlier, localised network of clustered interactions between a few closely-spaced enclosures (OA/Whitehouse 2013). The digging of the outer ditch defined not only the meeting space but also membership of a regional community, such as a lineage or clan. The years of seasonal labour (OA/Brown 1991) did not merely represent the community itself in action – it was the community in action. We suggest that the guardians of the outer enclosure lived in the C-shaped enclosures on the top of the site’s hill.

Figure 17 (a) Plan of Passo di Corvo (Whitehouse 2013, fig. 2.1G, amended by Lauren Woodard); (b) Plan of cobbled feature 215, Villeneuve-Tolosane

(source: Simonnet 1980, fig. 2.2a, amended by Lauren Woodard)

It is important to note that ceremonial events were not restricted to the congregation space at Passo di Corvo. There were many small-scale ceremonial acts and ritual performances – if few staged mortuary events – in the open areas in the inner enclosure, including feasting and special deposition of objects and body parts (for details, see OA/Tinè 1983; OA/Skeates 2000). The congregation area hosted the larger events, including ‘beating of the bounds’ (for explanation, see Section 1.3.2), regular deposition in the outer ditch and other less materially rich events (music, dancing, storytelling). While the excavation data show that similar practices occurred in the inner and outer enclosures at Passo di Corvo, the scale of the congregation and the intensity of the meetings between hundreds or even thousands of participants is what differentiated this site from all others on the Tavoliere.

4.3 Southern France

The later Neolithic in Southern France, designated as the Chassey group (Reference VaquerVaquer 1990), comprised a wide range of small settlement sites, with sizes between 0.07 ha and 5 ha, as well as enclosed sites attesting to both dwelling and display practices (OA/Phillips 1982). There were no sites as large as Passo di Corvo in Italy but two sites in the Middle Garonne Basin – St-Michel-du-Touche and Villeneuve-Tolosane – stand out relationally as the largest sites in the region – the latter over 400 times bigger than the smallest open site. Both sites show multi-period enclosure features, including ditches and palisades. Each successive extension of the enclosed area provided open areas for meetings. Cobbled surfaces (Fig. 17) sunk in pits provided the most frequent features at both sites and at other, smaller sites. We propose that these features were cooking pits where heated cobbles roasted entire skeletons, most often of cattle. A meat rather than a dairy profile of the Villeneuve cattle is consistent with feasting and the Garonne sites had less varied diets than those nearer the coast in Languedoc (OA/Herrscher & Le Bras-Goude 2010). Strontium isotopic mobility studies have shown that Languedoc residents were buried in the Garonne sites and vice versa (OA/Goude et al. 2012). The range of evidence is consistent with congregation sites at St. Michel and Villeneuve but without the regional focus of Passo di Corvo in the Tavoliere.

4.4 South-West Iberia

The use of aerial photography and geophysics has rarely had such a dramatic effect on a region’s archaeology as in South-West Iberia, with a new class of ditched enclosures of central importance since the 1990s. Although 60 per cent of these sites covered <1 ha (Reference Valera, In, Gibaja and ChambonValera 2012), Late Neolithic enclosures grew to 10 ha, with a further, Copper Age expansion to 110 ha. As with the apparently unenclosed megasite of Valencina de la Concepción (for extended discussion, see Section 4.5), excavations have demonstrated the multi-period, polyfocal nature of these major complexes. While a small Neolithic settlement preceded the Alcalar enclosure, cromlechs constituted the earliest monuments at both Perdigões and Porto Torrão. Multiple concentric ditches typified Perdigões (13 ditches: Fig. 18) and Monte da Contenda (17–19 ditches).

Figure 18 (a) Plan of Perdigões (source: Reference Valera, In, Gibaja and ChambonValera 2012, fig. 2.3/1, re-drawn by Lauren Woodard); (b) exotic exchange and non-local people buried at Perdigões: Areas: A – Estremadura (marble, limestone, people); B – Tagus and Sado estuaries and coastal Alentejo (marine shells); C – Estremoz – Borba – Vila Viçosa (marble, limestone); D – Tierra de Barros (pots made of schist-rich clay); E – Pico Centeno (variscite); F – Almaden (cinnabar); G – Jurassic silicified limestone outcrops, Betic Mountains; H – western North Africa (marine shells, ivory, people). As yet unsourced raw materials include gold, amber and rock crystal

(source: OA/Valera 2017, fig. 3.1, re-drawn by Lauren Woodard)

The rarity of banks and houses at most ditched enclosures and the high frequency of cut pit features meant that the most monumental features were often tombs, especially in the Copper Age. This spatial expansion of the mortuary domain was integrated with the burial of larger numbers of bodies/body parts with a higher proportion of exotic grave goods to create a new type of person-place, often interpreted as high-status. Sitting in a monumental landscape, the Perdigões enclosures were unique in having entrances constructed to face east to a valley containing over 100 megalithic tombs (Reference Valera, In, Gibaja and ChambonValera 2012).

There is an ongoing debate over the level of permanent settlement in ditched enclosures – if any dwelling at all to intensive occupation (Reference Hurtado, Valera and Shaw EvangelistaHurtado 2006). Given the importance of querns for food production, there is a striking similarity in their density at open sites, ditched enclosures and fortified sites (OA/Risch 2013). The focus of open areas for congregation at complexes with multiple ditches became greater, moving ever outwards with time. Larger-scale congregations could have been held in the outer enclosures at Marroquíes Bajos and Porto Torrão.

The congregational catchment of ditched enclosure communities can be studied through investigations of both human bones and objects. In the Middle Neolithic, a complex mortuary network distributed human bones across the landscape from a series of different cemeteries, all of which sent bones to the upland Bom Santo cave (OA/Carvalho et al. 2019). By the Copper Age at Perdigões, most of the large sample of sixty-nine individuals so far analysed for strontium isotopes came from outside the local area, with several persons matching signals from the Lisbon Peninsula, in contrast to the local origins of persons buried in dolmens (OA/Valera et al. 2020) (Fig. 18). Dental analysis shows North African influences in the distribution of some non-metric traits – possibly the result of genetic exchanges with North African groups (OA/Cunha 2015). This result is supported by the deposition of African elephant ivory in tholos tombs, often with Sicilian amber. The regular acquisition of stone and metals for Perdigões (Fig. 18) and Alcalar from sources up to 70 km away indicates that this was their likely congregational catchment zone.

4.5 Valencina de la Concepción

Of all the sites discussed in this Element, Valencina is the only one for which the core premise of congregation has been mooted (Reference García Sanjuán, Vargas Jiménez and Cáceres PuroGarcía Sanjuán et al. 2018, 280; Reference Martínez-Sevilla and Lozano RodríguezMartínez-Sevilla et al. 2020). The coastal site was located on the flat, fertile Aljarafe plateau, at 150 masl, overlooking the Guadalquivir valley and a coeval marine inlet. The AMS dates for Valencina show a long occupation from 3200 to 2300 cal BC (Reference García Sanjuán, Vargas Jiménez and Cáceres PuroGarcía Sanjuán et al. 2018), covering most of the Iberian Copper Age. It is also one of the largest complexes, at c. 450 ha, and one of the most intensively investigated sites in Spain, with nearly 130 excavation campaigns, starting in the 19th century (Reference Costa Caramé, Díaz-Zorita Bonilla, García Sanjuán and WheatleyCosta Caramé et al. 2010) (Fig. 19). The most prominent archaeological features in the present landscape are the impressive tholoi of La Pastora (Fig. 21) and Matarrubilla that once were part of a wider megalithic landscape that included mud-vaulted tholoi, the most impressive of which was Montelirio (OA/Fernández Flores et al. 2016). These funerary features were complemented by artificial caves (hypogaea) and pits with collective or individual burials, with both complete and partial bodies being common throughout the site. Last but not least are the numerous negative features, some of which showed clear evidence for production of metal and ivory, while others were more readily associated with domestic discard (Reference Costa Caramé, Díaz-Zorita Bonilla, García Sanjuán and WheatleyCosta Caramé et al. 2010). The sheer number of investigations conducted by almost as many different teams has led not only to great disparity in the level of systematisation and recording of the archaeological evidence but also to contrasting views about the nature of the site. For instance, Nocete Calvo et al. (OA/2008) have argued for Valencina as a state-level metallurgical centre – a view opposed by García Sanjuán and Murillo-Barroso (OA/2013). This lack of an integrated approach has plagued the understanding of Valencina for a long time; only in the last decade have huge strides been made to overcome the piecemeal utilisation of the available archaeological data, in which compelling interpretations of certain sectors/sites (e.g. Reference García Sanjuán, Cintas-Peña, Díaz-Zorita Bonilla, Müller and HinzGarcía Sanjuán et al. 2019) match holistic attempts at the understanding of the site (Reference García Sanjuán, Vargas Jiménez and Cáceres PuroGarcía Sanjuán et al. 2018, Reference Martínez-Sevilla and Lozano RodríguezMartínez-Sevilla et al. 2020). In Online Appendix II, we offer an assessment of the implications of the current views about Valencina, as well as a breakdown of the occupation in time and space into five phases in order to overcome definitional issues with dwelling and burial.

Figure 19 Site plan of Valencina de la Concepción, showing excavated areas

(source: Reference Costa Caramé, Díaz-Zorita Bonilla, García Sanjuán and WheatleyCosta Caramé et al. 2010, fig. 2.2, with additions by Lauren Woodard)

As with the other megasites we discuss, the question of why Valencina was so large has hardly been directly addressed. We propose that Valencina was a congregation site and its size was a consequence of both increasing place-value and internal development. The accumulation of cultural memory and place-based associations over nine centuries of occupation created a unique reputation for Valencina but not one without tensions. The multiplicity of domestic and mortuary foci suggests great dynamism and inter-focus contrasts in site development, leading to spatial displacement rather than ‘local’ transcendence of differences – all of which required a larger site area. Moreover, Valencina is a summary statement of the Iberian Copper Age settlement pattern containing dispersed dwelling and production areas, monumental tomb architecture, ditched enclosures and non-megalithic burials, only lacking the walled enclosures so typical of 3rd millennium Iberia. The incredible artefactual wealth of the site is not simply a result of long-term accumulation and random deposition but a further summary statement of key resources and the materialisation of wide exchange networks. Thus, the importance of dehesa forest management and land use in 3rd millennium Spain led to the crafting at Valencina of miniature images of pigs and acorns from ivory, while the prestigious rearing of equids (OA/Harrison, R. J. 1985) stimulated their fired clay representations and wide-ranging personal networks produced many exotic deposits. These are just a few aspects of a much wider and very diverse pattern that due to lack of space we cannot elaborate on here.

4.5.1 Scale

The size of Valencina, at 450 ha, has implications for the huge scale of social practices in an area not closely defined by a perimeter earthwork. The scale of occupation is indicated by the estimated number of 40,000 prehistoric features, based on extrapolation from the PP4-Montelirio sector excavations (Fig. 20). An as yet undefined proportion of these features can be related to congregational practices, of which a classic example is the compelling evidence for feasting in Structure 10.024 (Reference García SanjuánGarcía Sanjuán 2017). The unprecedented intensity of occupation should be integrated with the scale of consumption of certain artefacts and also the proliferation of special kinds of deposition. Currently there are several types of artefacts and materials in Valencina that are either most numerous or unique in comparison with other sites (Fig. 21). Not only is Valencina the sole site so far with both Asian and African ivory but also the quantity of ivory, at 8.8 kg, with 7.9 kg found in just two features (Montelirio tholos and PP4-Montelirio) far exceeded the amount at other sites (e.g. 1.7 kg at Perdigões). Local production of exquisitely crafted unique objects is attested and the site has provided the greatest number of gold (OA/Murillo-Barroso et al. 2015) and rock-crystal (OA/Morgado Rodriguez et al. 2016) artefacts of any Iberian Copper Age complex. The quantity of Sicilian amber far exceeded the usual low-level consumption, with only one site – Anta Grande do Zambujeiro – showing a comparable quantity of this exotic material (OA/Odriozola et al. 2019). These exotics attest not only to a special scale of consumption but also a huge reach in the scale of networking.

Figure 20 Detail of excavation plan of the La Pastora – Montelirio sector

(source: Reference García Sanjuán, Vargas Jiménez and Cáceres PuroGarcía Sanjuán et al. 2018, fig. 3.6: re-drawn by Lauren Woodard)

Figure 21 Exotic exchange at Valencina de la Concepción: A – amphibolite for querns; AF – North African ivory and ostrich eggshell; C – cinnabar; CS – calcareous sandstone building stone; F – flint (Milanos, Turón & Malaver types); FRA – flint, rock-crystal and actinolite from Baetic Mountains; G – granite for pounders and sandstone-shale building stone; GY – gypsum building stone; M – milonite; R – rock crystal, Albuquerque area; RH – rhyolite; S – salt; V – variscite; mining symbol – Rio Tinto and Aznacóllar copper mines;

(source: authors, re-drawn by Lauren Woodard)

It is through the totality of the evidence and the cumulative value of interactions that the uniqueness of Valencina’s scale of practices can be judged. It is in the concentration, combination and re-combination of varied practices – including the number of burials (probably exceeding 1,000 individuals), variety and quantity of exotics, variety of monument types and examples of unique craftsmanship – that makes the scale of Valencina stand out in Copper Age Iberia.

4.5.2 Temporality (Fig. 22 and Online Appendix II)

The groundbreaking cooperation that led to a comprehensive dating programme (Reference García Sanjuán, Vargas Jiménez and Cáceres PuroGarcía Sanjuán et al. 2018) underlines the cumulative duration of social practices at Valencina, matched at few other Iberian sites (Table 3). There was a strong contrast between the seeming permanence of the mortuary tholoi and the episodic utilisation of domestic areas, some even lacking clear residential traces. But the frequency of return visits to Valencina contributed to the narrative of an ever-changing regional population with material and emotional links to their congregation centre, turning the latter into a place more permanent than many small settlements.

Paradoxically, the precision of AMS dates for each site hinders a straightforward comparison of diachronic change, which is more readily achieved using 200-year time-slices (Fig. 22) (Online Appendix II). There is an inevitability about the dynamic, changing use of different parts of the site through the long occupation, with life starting with burial practices in at least two zones, finishing with an ancestral megalithic landscape with continued burial and sandwiching episodes of burial, production, permanent residence and seasonal visits of various durations. The upstanding tholos La Pastora (Fig. 23) encapsulated temporal relations linking the past (the inclusion in the roof of an altered capstone from another megalith), the present (ongoing mortuary usage) and the future (a spectacular but undated hoard of twenty-nine spearheads near the tholos (Reference García SanjuánGarcía Sanjuán 2017). The selection of different spatial foci for mortuary practices contrasted with the continuity of use of the central settlement area, with the different ‘local’ identities implicit in spatially disparate foci perhaps the main reason for the lack of an all-embracing perimeter enclosure.

Figure 22 Plans of the five phases of Valencina de la Conceptión

(source: authors, re-drawn by Lauren Woodard)

Figure 23 (a) photo of La Pastora tholos (source: authors); (b) reconstruction of the grindstone-breaking and burning ritual

(source: Lauren Woodard)

Valencina must very often have stood out among its contemporary counterparts with its unique use of space, which in its congregational function accommodated both continuity (settlement foci) and change (from collective to individual burial). Some thirty to forty generations of people would have built up a unique combination of architectural foci and cumulative social memory through repeated acts of production and deposition.

4.6 Conclusions

The repeated occupation, deposition and burial at Valencina de la Concepción created a patchwork of rich cultural memories, which, by the latest deposition, stretched back almost a millennium. This summary of site and monument forms that typified the Iberian Copper Age made Valencina a unique congregation place to which thousands of visitors returned repeatedly, bringing some of the rarest as well as the most common objects of the age for deposition in special and quotidian places. The scale of performances and celebrations of the annual meeting made Valencina one of the most famous sites in Western Europe, with its reputation spreading out to many regions and thousands of sites.

5.1 Introduction

Few European prehistorians would dispute that Bronze Age networks connected more people over longer distances with more objects – especially metal – than in previous times. However, there is far less agreement over the political and military consequences of these changes (OA/Kienlin 2020; Reference Hansen and KrauseHansen & Krause 2018). A focal point of development came in the Late Bronze Age, starting c. 1400 cal BC, with the construction of over 1,000 hillforts from Central Germany to the Carpathian Basin (Reference Hansen and KrauseHansen & Krause 2018) accompanying major innovations in weaponry and armour. Fortifications create a major challenge to what Keeley (OA/1996) has called ‘a pacified prehistory’ and challenge us to explain what was one of the greatest transformations of the Bronze Age.

For Reference Keeley, Fontana and QuickKeeley et al. (2007, 79), ‘the symbolism of fortifications was always predicated on their military functions’. Although not all enclosing features (or ‘enceintes’) were defensive in nature, enclosures could still be defensive even if they lacked the three key traits of a fortification – V-sectioned ditches, complex gateways and bastions (Reference Keeley, Fontana and QuickKeeley et al. 2007). Defences protected a wide range of the elements of everyday life, whether resources and possessions, homes and special places such as shrines, as well as transcending their military functions to increase status and monumental performance.

Located in the South-Eastern part of the Middle Danube Basin (Fig. 24), the lowland region of the Banat reveals the greatest concentration of large enclosed/fortified sites, termed ‘megaforts’ (Reference Molloy, Jovanović and BruyèreMolloy et al. 2020, 293), in Bronze Age Europe, with the highest number of LBA gold finds in the Basin (OA/Mozsolics 1973) found in the key zone of the Lower Mureş/Maros valley. In the preceding Middle Bronze Age, the principal form of central place was the tell, whose elevated occupation levels were often defended with encircling ditches rather than banks (Reference GogâltanGogâltan 2017). The main tell in the region has been identified as Pecica, which, at its peak, was linked to the distribution of horses and metalwork throughout the Carpathian Basin (OA/Nicodemus & O’Shea 2019) (Fig. 24/Site 10). Although tells founded in the MBII or III phases are not known in the Lower Mureş, the MBA Phase I tells formed antecedent monuments still visible in the landscape, playing significant roles in several LBA megaforts (Table 4). The lack of chronological overlap between MBA fortified tells and LBA megaforts does not preclude a shared habitus of living in a clearly delineated, highly structured enclosed space.

Figure 24 Map of Late Bronze Age sites in South-Eastern part of the Middle Danube Basin, with copper sources

(sources: Reference Molloy, Jovanović and BruyèreMolloy et al. 2020, fig. 1.1; OA/Mareş 2002, Harta 1: re-drawn by Lauren Woodard)

Reference Molloy, Jovanović and BruyèreMolloy et al. (2020, fig. 1.1 & table 1.1) have identified a site hierarchy based on scale and size, design and distribution, with seven size-based ranks. Even if the ranking is over-elaborate, it is clear that each megafort in the top rank organised a network of smaller enclosed forts and even smaller unenclosed settlements (Fig. 24/Sites 6–8). Molloy et al. interpret this settlement pattern as a series of multi-local societies under a common political framework, with the construction and maintenance of megaforts a performative form of power. ‘Enclosures are a physical embodiment of a recognisable regime that resourced large-scale political ideologies manifested through similar settlement features … in a densely settled landscape’ (Reference Molloy, Jovanović and BruyèreMolloy et al. 2020, 299). But LBA communities structured not only their own fortified settlements. The presence of a linear ditch system almost 10 km in length linking four LBA forts between Variaş and Satchinez shows the landscape scale of land division and territorial control (OA/Dorogostaisky & Hegyi 2017, who cite several other cognate systems) (here, Fig. 25). It was not only at the site level of change from tells to megaforts that LBA social transformations were played out but at the wider landscape scale as well.

Figure 25 (a) Plan of the Variaş–Satchinez linear ditch system (source: Dorogostaisky & Hegyi 2017, fig. 2.3, re-drawn by Lauren Woodard); (b) Interpretation of geophysical plot, Sântana

(source: Reference Gogâltan and SavaGogâltan & Sava 2010, fig. 2.6, re-drawn by Lauren Woodard)

5.2 Important Megaforts

The exceptional size of Corneşti-Iarcuri singles the complex out for special discussion (see Section 5.3). Four other well-investigated sites provide a representative selection of the fortified examples of the seventy currently known LBA sites in the Banat (Reference Molloy, Jovanović and BruyèreMolloy et al. 2020, fig. 1.1) (here, Fig. 24) – Csanadpalota (400 ha), Idjoš-Gradište (200 ha), Sântana (180 ha) and Munar (15 ha) (for details, see Table 4). All of these sites, which are dated to 1400–900 cal BC, shared a common reliance on both natural and cultural antecedent landscape features (OA/Zvelebil & Beneš 1997). Streams flowed across the enclosed areas, actively influencing the layout of the defensive system, as at Sântana (Fig. 25). The provision of fresh water for animals and humans may ultimately have been compromised by waste disposal. In three cases, the forts were built next to, or even included, earlier tells – a Late Neolithic tell at Idjoš and MBA tells in the Munar complex and near Sântana. Four barrows were included in the last-named’s enclosure, with other barrows outside, while a network of forty barrows forming three linear groupings along palaeo-channels probably pre-dated the Idjoš enclosures (Fig. 26), as did a flat inhumation cemetery 1 km distant. These ancestral links to nature and earlier settlers tied the identities of the fort-builders into their local landscapes, legitimising their presence at key nodes in the settlement network.

Figure 26 (a) Interpretative plan of the geophysical plots, Idjoš – Gradište complex (source: Reference Molloy, Jovanović and BruyèreMolloy et al. 2020, fig. 2.3B, re-drawn by Lauren Woodard); (b) general plan of Corneşti–Iarcuri: red areas – excavated

(OA/Szentmiklosi et al. 2016, fig. 1.1, re-drawn by Lauren Woodard)

An important characteristic of these sites is the marked variability in enclosure design. The most importance difference distinguishes the majority of sites, with their concentric enceintes (e.g., Sântana: Fig. 25), from the Idjoš example of agglutinative design, where four separate enclosures have been constructed next to one another near the small mound termed the ‘citadel’ (OA/Marić et al. 2016) (Fig. 26). These designs suggest contrasts not only in temporality, with the obvious sequence of the concentric enceintes and the far less clear sequence of the Idjoš enclosures, but also in social relations – building on traditional, ancestral structures in the former, with the possibility of contrasting identities in each enclosure with the latter. The diversity of combination of the same four basic elements – banks (with or without casemate construction), ditches (of varying form), berms and palisades – varied from site to site and sometimes within sites (e.g., the different forms of bank construction in the three banks at Sântana). Special features also occurred, such as the 2.5 km long linear ditch in the centre of the huge Csanadpalota enclosure – reminiscent of the Variaş–Satchinez linear ditch system – or the formalised entrance to the Idjoš citadel, access to which required crossing six ditches. The form of the ditches meant that defence was prioritised at some sites (e.g., the V-sectioned ditch at Sântana Enclosure III) but not at others (V-sectioned ditches were not encountered at Csanadpalota). At Idjoš, the form of ditch 2A, dug in front of an entrance, changed from V-sectioned, so defensive, to a U-shaped profile and finally a shallow, flat-bottomed form (Reference Molloy, Jovanović and BruyèreMolloy et al. 2020).

The widespread, indeed essential, use of geophysical investigation at these large sites has also produced widely varying results. Reference Molloy, Jovanović and BruyèreMolloy et al.’s (2020) reply to the question of why there were so few house-shaped anomalies at Idjoš was that the architectural design itself left few traces, but this is contradicted at Sântana Enclosure I, with its cluster of impressive rectangular houses in one 60 m x 40 m complex. A series of ditches in Enclosure II at the same site has been interpreted as a delineation of residential areas. This would suggest that the absence of building-type geophysical anomalies may indeed indicate evidence of absence and not the converse. Little evidence for craft production has yet been found at these sites, although slingshot production may have been a local skill highly developed at Sântana for resistance to the putative siege postulated at this site. Rather than military preparedness, place-marks in ditches and pits are indicative of special deposition at all of these sites, as in the long life-cycle of certain pits at Idjoš and the deposition of the bones of game animals, perhaps from elite hunting.

Feasting has been claimed at Csanadpalota, where a group of five pits shows evidence for the deposition of feasting remains (e.g., Pit 474, with the bones indicating the consumption of meat weighing 600 kg). The overall faunal sample shows a mix of low-, medium- and high-quality meat cuts for all domestic species and red deer, suggesting both feasting and ‘normal’ domestic consumption may have relied on all meat qualities (Reference KassabaumKassabaum 2019). Large-scale megasite feasting may well have led to different middening practices, with pit middens exhibiting a slower, gradual build-up of feasting remains but still within the year, while open-air middens would have received feasting debris more rapidly. This feasting and other evidence suggested to Reference Szeverényi, Priskin, Czukor, Kniesel, Dal Corso and KirleisSzeverényi et al. (2015, 101) the interpretation of Csanadpalota as ‘a fortified ritual centre where a large community gathered from time to time to carry out various rituals’ – a rather more positive conclusion than Reference Molloy, Jovanović and BruyèreMolloy et al’s (2020, 310): ‘we would struggle to rule (Idjoš) out as a venue for gatherings or events’.

There can be little doubt of the necessity of large numbers of people for the construction of massive dump ramparts at these sites. Estimates for the 3.63 km long Sântana Enclosure III rampart (Fig. 25) suggest a requirement of over 237,000m³ of earth, mostly sourced from the excavation of the adjoining ditch (Reference Gogâltan and SavaGogâltan & Sava 2010). Calculations using the same methods as discussed for Corneşti (Section 5.3.3 and Online Appendix III) but using the smallest estimated workforce show the requirement of between four and twenty construction seasons for the creation of all three ramparts. The scale of the gathering for Sântana and the other sites indicates political mobilisation at the regional level and hints at the principal function of the largest megafort of all.

5.3 The Corneşti–Iarcuri megasite

Corneşti is the largest megafort in the Banat by a factor of four. It is, indeed, the largest known Bronze Age site in Europe, with a total perimeter length from the four concentric ramparts of 33 km and an enclosed area of 1,764 ha, making it four times bigger than the Valencina de la Concepción Copper Age megasite discussed in Section 4 (Reference Szentmiklosi, Heeb and HeebSzentmiklosi et al. 2011; Reference Heeb, Szentmiklosi, Bălărie, Heeb, Szentmiklosi, Krause and WemhoffHeeb et al. 2017; Reference Krause, Szentmiklosi and HeebKrause et al. 2019) (here Fig. 26). While ramparts I and II are visible in the modern landscape, with heights of up to 4 m (Fig. 27), ramparts III and IV are far less visible and may have been lower than the inner banks in the past. Research at Corneşti since 2007 has relied on the three familiar techniques of investigation, supplemented by palaeo-environmental studies (OA/Nykamp et al. 2015; OA/Sherwood et al. 2013). However, the size of the complex is daunting and the absolute dating of the sequence of rampart construction within the Late Bronze Age and Early Iron Age remains uncertain.

Figure 27 (a) Cross-section 2 of rampart and ditches, Enclosure I, Corneşti–Iarcuri (source: Reference Heeb, Szentmiklosi, Bălărie, Heeb, Szentmiklosi, Krause and WemhoffHeeb et al. 2017, fig. 2.2, re-drawn by Lauren Woodard); (b)–(c) two models of the work study, Corneşti–Iarcuri (authors)

5.4 Conclusions

A decade ago, Reference Szentmiklosi, Heeb and HeebSzentmiklosi et al. (2011) maintained that the population concentrated in Corneşti was insufficient to support an urban function, while there were so many gateways in the ramparts that it could not have been a defensive structure. After much new research, both conclusions remain valid: the site seems best considered as a trans-regional centre in a network of major fortified sites – a ‘megasite’ (Reference Heeb, Lehmpful, Szentmiklosi, Hansen and KrauseHeeb et al. 2018, 395) with an emphasis on overt economic power and implicitly symbolised military power. If a military elite did indeed organise the construction of the ramparts at Corneşti, there has been little materialisation of their status, other than through the ramparts and ditches themselves. The paucity of permanent occupation features matches the criteria for its principal function as a place of congregation for large groups of people from the already large communities in sites such as Csanadpalota, Idjoš and Sântana, while this proposal is also supported by the super-abundance of open space in all four enclosures.

6.1 Introduction

The European Iron Age (800/700 BC–1 BC) lasted as long as the period when Trypillian communities were constructing megasites, or shorter than the total duration of the Valencina de la Concepción occupation. Yet the strongly regional series of dynamic changes in settlement and material culture underline the rapid pace of cultural change in the first millennium BC (Reference Moore, Armada, Moore and ArmadaMoore & Armada 2011). In the Online Appendix IV, we present five regional trajectories to give a flavour of the varied settlement histories that gave rise to Iron Age megasites. These trajectories share the unifying feature that the Iron Age was ‘still principally a world of the common farmer’ (Reference Danielisová and Fernández-GötzDanielisová & Fernández-Götz 2015, 9), for whom the vast majority of settlements constituted small farmsteads comprising populations of fewer than 100 persons, each producing their own pottery, iron tools and ornaments (Reference Haselgrove, Rebay-Salisbury and WellsHaselgrove et al. 2018). These views have been generalised into the ‘segmentary society’ model – the default settlement unit in the European Iron Age and the polar opposite of the megasite.

The megasites that can be identified in both the Early (EIA) and the Late (LIA) Iron Age (Fig. 28; Table 5) share the main characteristics of low-density dwelling with large areas of open space within a well-defined perimeter.

Figure 28 Distribution of (upper) Early Iron Age Fürstensitze and (lower) Late Iron Age oppida

(source: Reference Fernández-GötzFernández-Götz 2018, figs. 1.1 and 2.7, re-drawn by Lauren Woodard)

The monumentality of the EIA centres was another principal characteristic (Fig. 29), embodying both the idea of construction labour as potlatch and also the rhetoric of stability and permanence despite architectural variations through time. Information on Early Iron Age exchange networks, including building materials (Fig. 30), similarities in hillfort planning suggesting shared design and the diet of stock from isotopic studies all combined to suggest a congregational catchment of 60 km for the Heuneburg.

Figure 29 (a) Plan of the Heuneburg complex (source: OA/Fernández-Götz & Ralston 2017, fig. 2.1: amended by Lauren Woodard); (b) Plan of the Mont Lassois complex

(source: Chaume 2020, fig. 21.1: with additions by Lauren Woodard)

Figure 30 Map of the Heuneburg territory

(source:OA/ Steffen 2012, Abb. 94, amended by Lauren Woodard)

The largest LIA sites operated at a far larger scale than EIA centres, with landscape features incorporated into sites such as Heidengraben (Fig. 31) or areas of 3,000 ha between oppida forming coherent territories (Fig. 31). Internal oppidum space was highly structured, with feasting a major pre-occupation.

Figure 31 (a) plan of Heidengraben (source: Reference Ade, Fernández-Götz, Rademacher, Stegmaier and WillmyAde et al. 2013, Abb. 46, re-drawn by Lauren Woodard); (b) plan of the Corent – Gondole - Gergovie complex

(source: Reference PouxPoux 2012, map, p. 259, with additions by Lauren Woodard)

While the regions of the European Iron Age do not constitute unitary phenomena, the megasites are linked by their shared polyfocal development, with a marked increase in scale of occupation from the EIA to the LIA. We now turn to a classic example of a polyfocal Iron Age megasite – the largest enclosed site in Europe.

6.2 The Bil’sk polyfocal megasite

6.2.1 Introduction

Bil’sk is currently the largest known fortification from the 1st millennium BC in Europe (Fig. 32). It is located in the forest-steppe zone of modern Ukraine, in the watershed between the rivers Vorskla and Suhaja Grun – left-bank tributaries of the river Dnieper (Fig. 33). The site’s ceramic-based chronology, including Greek imports, show occupation in the 8th–4th centuries BC, although earlier and later interventions are also known. Palaeo- environmental studies suggest little difference in vegetation cover even at the time of most intensive occupation, although the climate was cooler and wetter (OA/Сорокіна et al. 2014). The first excavations took place in 1906 and intermittent investigations continue to this very day. The exact area of the fortification is contentious but an enclosed area of ca. 5,000 ha is unparalleled in the contemporary world. Excavation of just 0.2 per cent of the enclosed area is supplemented by remote sensing (geophysical plots (OA/Орлюк et al. 2016) and/or GIS-reconstructions (Reference DaraganDaragan 2020)) of large parts of the site. In addition to the main rampart, there are three hillforts at Bil’sk – all structurally linked to the main wall.

Figure 32 Plan of Bil’sk: red - settlement zones; green - barrow zones (source: John Chapman, from information in Reference Скорий, Білозор, Супруненко and КулатоваСкорий et al. (2019), Ris 299; Reference ШапордаШапорда О. (2017), Ris. 1–2; Reference Johnson, Pankova and SimpsonJohnson (2020), fig. 2.3, re-drawn by Lauren Woodard)

Figure 33 (a) Map of region, with Greek colonies, Bil’sk and Dnieper right-bank hillforts (source: Reference Johnson, Pankova and SimpsonJohnson 2020, fig. 2.2, re-drawn by Lauren Woodard); (b) distribution of hillforts in inner zone of Bil’sk

(source:  OA/Білинський (2018), Ris. 10, re-drawn by Lauren Woodard)

6.2.2 Site Formation

Discussions of site formation have no tradition in Eastern European archaeology. Bil’sk is no exception and the complex site taphonomy is poorly understood.

The most visible parts of Bil’sk are the ramparts topped with palisades, reaching up to 9 m in height, with ditches of up to 6 m in depth (Reference DaraganDaragan 2020, figs. 3.4–3.5 for palisade reconstructions). The other main archaeological features are zol’niki, barrows and maidans. ‘Zol’niki’ typified the Late Bronze Age–Early Iron Age along the Dnieper. A zol’nik is an ash deposit that, before ploughing, may have reached more that a metre in height (now 0.50 m), with a circular/oval shape varying between 20 and 100 m in diameter (OA/Дараган & Свойский 2018) (Fig. 34). The main archaeological contexts in zol’niki are cut features such as pits and semi-subterranean dwellings, above-ground dwellings and workshops. Most of the archaeological material is found within/under the fill of these cut features but also between the contexts in cultural layers. Domestic hearths, kilns, furnaces and altars are also found in the cut structures but can be free-standing (e.g. Reference КоростКорост 2016, 519). Traces of occupation are found between zol’niki, too, but have been rarely investigated. There is a contrast between the Western hillfort (WHF) with at least fifty-four zol’niki (OA/Шрамко 2012) and the Eastern hillfort (EHF) with none (Reference КоростКорост 2016: 88) but containing similar cut features without ash-deposits.

Figure 34 (a) plan of zol’nik, with pit and adjacent building (source: OA/Шрамко & Задніков (2019), Ris 2); (b) plan and cross-section of zolnik found beneath a barrow (source: Reference КоростКорост, (2016), p. 436); (c) plan of barrows and majdans outside Bil’sk ramparts

(source: OA/Махортых (2013), Ris. 2) (all amended by Lauren Woodard)

The barrows are straightforward earthen mounds, in some cases covering wooden chambers built in large pits (for reconstructions, Reference DaraganDaragan 2020, fig. 3.6) (Fig. 34). Currently barrow size varies between 0.15–1.85 m in height, and 15–100 m in diameter (Reference ШапордаШапорда 2017) but with loss of height. A total of 61 per cent of the 103 excavated burials contained weapons of some sort (OA/Махортых & Ролле 2011), suggesting diverse identities in the buried population. Zol’niki have also been discovered under some mounds (Reference КоростКорост 2016: 359, 362, 434–6).

Maidans are a contentious feature class whose lack of burials suggests they were not ploughed-out barrows. They lie midway between earthworks and barrows and are best described as ring-mounds of unclear function (Reference КоростКорост 2016: 550). The maximum height is 4 m and diameter 330 m, with the majority varying between 0.4–1.5 m in height and 30–135 m in diameter (Reference ШапордаШапорда 2017) (Fig. 34).

6.2.4 Scale

In an area of forest-steppe of 120,000 km² on the left bank of the Dnieper, there were 138 mostly small hillforts (Fig. 33), dated within the 7th–4th centuries BC, of which two (Bilsk and the much smaller Knishinvka) are dated to the 7th century BC. Although only one hillfort (Basivka) is over 100 ha, on the right bank of the Dnieper, 300 km away, there are nine more large hillforts enclosing between 100 ha and 700 ha (Fig. 33). While there were precedents for huge sites, Bil’sk’s size was unprecedented. Hillfort function has been related to size, differentiated into small temporary shelters (1–3 ha), fortified settlements of relatively small tribal or territorial communities (3–10 ha: 250–500 people/ha), fortified settlements with citadels and/or suburbs (11–55 ha: 200 people/ha) and massive tribal agglomerations with short occupations (over 55 ha), some of which developed central roles in trade, handicrafts and administration and warfare, which put them on the road to local ‘urbanism’ (Reference КоростКорост 2016: 199–201).

This conflation of function, size and urbanism means that only Bil’sk is considered a ‘city’. If Bil’sk was to be ranked by size of built-up area among the largest 100 modern English cities, it would come in at forty-second place – well ahead of Oxford and York – and, based on modern population densities, with a population of just under 180,000. The maximum Ukrainian calculation for Bil’sk reaches 80,000 but 50,000 is usually quoted (Reference КоростКорост 2016: 132). Every large hillfort is seen as a centre of its own territory, which included settlements, smaller hillforts and barrows (cf. Late Hallstatt Fürstensitze: see Online Appendix IV) (Fig. 36). However, Bil’sk stands out from other sites, with its 50 km² of enclosed area as large as the entire ‘functional zone’ of some small hillforts and its functional zone as large as an entire micro-regional system (Reference БойкоБойко 2017: 31) (Fig. 36).

Figure 35 Size of interior sites within Main Rampart, Bil’sk

(source: authors)

Figure 36 (upper) settlement clusters in the Bil’sk region; (lower) Map of Thiessen polygons for Iron Age ‘territories’

(source: Бойко (2017), pp. 21–2: re-drawn by Lauren Woodard)

We maintain that what differentiated Bil’sk from other hillforts was its political centrality, requiring large-scale seasonal congregations involving representatives from all its many communities, in turn necessitating a vast and impressive enclosed space.

6.2.7 Monumentality

The size of Bil’sk is most clearly characterised by its huge ramparts (Fig. 37). But how were they built and over what period of time? We reject Gritsyuk’s (OA/Грицюк 2004) calculations, which over-rely on a level of standardisation not recognisable in the published cross-sections, leading to the incorrect assumption of identical construction methods over the total length of the main rampart. It is difficult to relate Reference Johnson, Pankova and SimpsonJohnson’s (2020: 208–9) work study of the Bil’sk ramparts to this study since he does not make explicit which rampart cross-section(s) he has has used in his calculations.

Figure 37 Cross-sections of Bil’sk ramparts: (a) West hillfort; (b) East hillfort; (c) main rampart

(source: Reference DaraganDaragan 2020, fig. 3.4, re-drawn by Lauren Woodard)

The aim of our work study was to make structured comparisons between the construction of three ramparts – those of the WHF, the EHF and the MR – in what appeared to be a continuous building programme that was intensified from the 6th century BC. The study is divided into two parts: a calculation of the number of person-years taken to build the three entities using four models (Table 6 & Fig. 38) and an assessment of overall construction time (Fig. 39).

Table 6 Four models for the construction of the Bil’sk ramparts

Model	Work-Force	No. of Days p.a.	Local Workers	Visiting Workers	Tools and Containers	Food and Drink
Congregation	5,000	20	–	5,000	Visitors	Bil’sk + basics; visitors + feasting
Corvée (A)	1,000	50	500	500	Each supply own	Bil’sk + food/drink for all
Corvée (B)	2,000	50	1,000	1,000	Each supply own	Bil’sk + food/drink for all
Part-time	1,000	70	500	500	Each supply own	Bil’sk + food/drink for all
Full-time	1,000	140	1 000	–	Bil’sk provides all	Bil’sk provides all

Figure 38 Three models for the construction of the Bil’sk hillforts and ramparts

(source: authors)

Figure 39 Three models of construction time for (a) main rampart, (b) West hillfort and (c) East hillfort, Bil’sk

(source: authors)

It is proposed that local full-time labour was used to make Phase 1 of the WHF rampart in less than one year as a symbol of local identity, power and self-sufficiency (Fig. 39). Thereafter, pottery dating suggests gaps of several decades between the construction of each successive Phase, which lasted from the late 7th into the 6th century BC. For the same reasons, the same full-time labour model is proposed for the construction of the EHF, with the Phase 1 rampart laid out in six months and only the Phase 4 rampart taking more than one year. By contrast, the construction of the MR was a symbol of the identity and political supremacy of a much larger regional group, requiring the contribution of labour from most, if not all, members of the Bil’sk congregation over many years. There was time for a generational gap of c. fifteen years between the constructions of each Phase of the MR, which covered the entire 6th century BC. This modelling contradicts the possibility of rapid, massive construction in a few years, arguing against a defensive function and in favour of community identity and symbolic power for all three entities.

Timber for the palisades atop the ramparts of the WHF and EHF could have been cut from the forests inside the hillforts. The same may have been true for the forested area of 400 ha required for the MR palisade. The total earth estimated for all Phases of the WHF combined was over 300,000 m³ (Reference DaraganDaragan 2020) – perhaps seven times the earth required for the MR. These daunting figures underline the monumentalisation of resources consumed as much as labour, indicating the potlaching of labour in the context of an increasingly impressive monument.

In contrast to the Heuneburg complex (see Online Appendix IV), there is a remarkable divergence between the overt political statement of the majestic ramparts and the fairly flimsy domestic architecture inside the Bil’sk complex (Fig. 32). This contrast tells us something about the type of occupation – e.g., political elites organising a regional if not inter-regional congregation for the purposes of social reproduction, rather than a single paramount chief or a succession of charismatic leaders maintaining the might of Bil’sk for three centuries.

6.2.8 Performance

If there is one site that highlights the importance of the formalization of space allowing processions, burials, depositions, chanting, feasts with copious wine, games and competitions – perhaps involving horses – on a grand scale, it is Bil’sk. Imagine the awe felt by a visitor from small sites and hillforts at seeing Bil’sk for the first time!

Formalised performance was central to the mortuary domain, with its overwhelming number of the Bil’sk barrows. The 130 barrows to the north-west (OA/Махортых 2012) and the ninety-nine barrows, fifteen maidans and eighteen barrow-like mounds to the west (Reference ШапордаШапорда 2017) make the hypothesis of over 1,000 monuments very plausible. The orchestrated ceremonialism of traditional burial practices is highlighted by the formal planning of thirty-two barrows arranged in four rows each of eight barrows of decreasing size and a matching row of eight maidans (Fig. 34). This layout channelled movement, affecting perception and reinforcing staged, repeated processions. With time, the steadily increasing number of barrows would have resulted in a multifocal ‘stage’, where public mourning, remembrance and celebrations would have been unparalleled in the contemporary world.

Performances varied between the different components of the non-mortuary domain, including the many sacrificial deposits and the sanctuaries involving calendrical and astronomical observations. The opening ‘horns’ of the maidans (Fig. 34) channelled participants into the central depositional area, while the scale of middening forming zol’niki could reach the size of a football pitch and must have involved many farmsteads. The linear feature in the middle of the WHF (OA/Дараган & Свойский, 2018) (Fig. 40) makes sense as a processional route from gate to gate, which could also have facilitated the movement of materials between zol’niki. Processions around the EHF, the WHF and segments of the MR would have formed important, regular calendrical events. These graded sequences of performance – from local farmstead to site-wide spectacle – created the basic structure of everyday life as well as setting the scene for significant congregational rituals.

Figure 40 (a) Western hillfort processional way (source; Дараган & Свойский (2018), Ris. 19, with additions by Lauren Woodard); (b) geological map of surrounds of Bil’sk, with S^r pattern

(source: Reference Ventresca Miller, Johnson and MakhortykhVentresca-Miller et al. 2019, fig. 2.1, re-drawn by Lauren Woodard)

6.3 Conclusions

Bil’sk was a colossus even by the scale of other European prehistoric megasites. Reference Taylor, Havlicek, Beckwith, Pankova and SimpsonTaylor et al. (2020: 622) put this in context: ‘We need to simply accept that the Scythian social formation by the 5th and 4th centuries BC represented something far more militarily, and even politically, powerful than all the Greek city states put together.’ For Bil’sk, this translates into a massive congregation place, whose ramparts defined open areas into which all of the other megasites discussed in Sections 2–5 would have fitted and which demonstrated the widest conceivable range of practices completed by a highly heterogeneous population. If this volume required a single exemplar of the congregational function of a megasite, we need go no further than the Ukrainian forest-steppe in the Iron Age.

7.1 Introduction

In her recent book about capitalism, Reference MazzucatoMariana Mazzucato (2021) makes the basic distinction between profit-oriented capitalist systems and mission-oriented projects, with the US moon landing project of the 1960s the classic example of the latter. While Mazzucato identifies President J. F. Kennedy as the driving force behind the moon mission, the difference that he made to the mission once it was under way steadily decreased, insofar as its success depended on the assent of thousands of dividuals and hundreds of companies. One of the reasons that Mazzucato makes so much of the moon mission is that there have been few other shining examples of goal-oriented successes in the late 20th and early 21st century, although global heating provides a huge challenge. Can we learn anything about prehistory from the moon-landing?

In prehistory, social groups tended to be goal-oriented, viz., mission-oriented, with as many people as possible contributing to wider social goals such as the construction of a megalith or the excavation of a copper mine. In this sense, to what extent were the congregation places that we have been discussing in this Element the product of goal-oriented systems? This description fits well with three megasites defined by looking inwards from the outside – Nebelivka, with its clearly delineated plan based upon a perimeter ditch and concentric house circuits, and the two megasites defined by massive ramparts – Corneşti and Bil’sk. The absence of perimeter ditches at both Alsónyék and Valencina conveys a different sense of sites defined by looking outwards from a crowded interior, in which the cumulative growth of both dwelling and mortuary features provided a forward momentum that is as close to a goal orientation as one can find there.

This answer underlines the importance of a comparative approach that takes into account the differences as well as the similarities in our sample of megasites. In this final section, we seek to draw together and compare the key elements of megasites in prehistoric Europe. We conclude by examining some basic ideas surrounding the linkages between megasites and their antecedent backgrounds as well as their internal biographies – their birth, life and death.

7.2 Why Megasites?

Why did megasites occur at all in prehistoric Europe and why were they so rare? If megasites were so vital to functioning social orders, why were there not more examples? Even with dating revisions to known monuments (e.g., the dating of the Dartmoor reaves to the Bronze Age: OA/Fleming 1988), the recognition of ‘natural’ features as the product of human action (e.g., the Potterne MBA complex: OA/Lawson 2000) or the expansion of LIDAR reconnaissance (e.g., the discovery of the full extent of the Greater Angkor system: OA/Evans et al. 2007), a sudden influx of hitherto overlooked megasites across Europe seems improbable. Were the megasites discussed here so different that each evolved by chance in a totally unrelated way in different millennia? Or were there common threads linking this class of site? Do they even constitute a separate ‘class of site’?

Let us begin to answer these questions in reverse order. The global settlement research of Reference FletcherRoland Fletcher (2019) and Kirrily White (OA/2022) demonstrates that ‘Anomalous Great sites’ (viz., megasites) occurred from the 5^th millennium cal BC onwards in every inhabited continent, sharing a wide range of temporalities and practices. The long-term Holocene settlement context of seasonal nucleation at meeting sites and dispersion into the basic unit of settlement – the forager camp, segmented village or farmstead – shows the importance of seasonal aggregation as an evolutionarily stable strategy. Figure 41 shows the wide variety of megasite sizes in any one period, while underlining the gradual increase in maximum size of megasites over time. But megasites took the process of seasonal meeting several steps further into territory that had never occurred before and had to be imagined before it could develop (Reference GaydarskaGaydarska 2020: 36; cf. OA/Anderson 1991). This is an important parallel between megasites and the moon mission. This imagined community became far larger than any existing settlement or even any known meeting place, even though it began life as something more ‘normal’. It relied on a balance of positive inter-settlement interactions over negative results, with the deposition of exotic materials and objects presencing a massive exchange network. It showed a commitment to something grand and spectacular while at the same time constituting a summary statement of all the important elements of the wider social network. These achievements were shared by all of the megasites in our study; we believe that, although some, if not all, aspects were found individually on other classes of site, the integration of all these developments were unique to megasites and characterises them as a separate class of site.

Figure 41 Sizes of selected megasites by period

(source: authors, re-drawn by Lauren Woodard)

The rise of rare megasites would appear to be a classic opportunity to invoke the ‘charismatic leader’ (OA/Jones & Kautz 1981) – the prehistoric J. F. Kennedy who could integrate all of the disparate elements of the social network to form a megasite widely supported by all elements of society. However, since this explanation is unfalsifiable sensu Popper (OA/1959), even in the absence of spectacular single burials, we shall not consider it further. It is a lost-cause form of explanation, when all other explanations have failed.

Did these imagined communities arise sporadically for different reasons in different time-place contexts? ‘Yes’ and ‘no’. Yes, in that we would expect a set of common underlying factors causing the situation to which a megasite was an appropriate answer – perhaps concerning network issues such as access to exotics and rare skills. No, in that the response to the situation would have been shaped by different antecedent social structures and practices, not least the degree of political centralisation and the regional heritage of different classes of sites and monuments. Clearly, the change from defended tells to forts in the Carpathian Basin at the MBA–LBA transition (Reference GogâltanGogâltan 2017) invoked very different antecedent structures and practices from the development of a mortuary congregation at Alsónyék.

However, we can observe three general aspects of the relationships between megasites and their antecedent contexts. First, there were elements of the wider landscape that were incorporated as pre-existing monuments in the nascent megasite – monuments that acted as a familiarising context for the wider region. Secondly, the ‘final’, grandest site form – what attracted archaeologists to the site as a megasite in the first place – was the product of the long development of a megasite’s biography, which often started out much smaller and only grew into its final iteration after centuries of punctuated development and a cumulative increase in place-value (e.g., the concentric expansion of multi-ring enclosures). The positive experiences of past visitors at these places made it more likely that the advantages of scale would be further appreciated into the future. Thirdly, and perhaps most importantly, an over-riding characteristic of congregation sites is the way in which they provided a summary statement of all of the major features of the contemporary landscape – whether banks, ditches, pits, ramparts, monumental tombs and dwelling features – as well as a summary of all the key materials and objects in their regional or inter-regional exchange networks. We suggest that, in these ways, the cumulative transfer of the known into basic elements of the new and ‘unfamiliar’ presenced a huge cast of absent people, places and things and made the unimaginable possible. We now turn to examples of these common elements in the megasites we have discussed, beginning with antecedent features.

7.3 Antecedent Sites and Monuments

It would seem that, since all of the five megasites discussed in detail here derived from different cultural and social contexts, they would have developed out of contrasting antecedent social structures and practices. However, all of the megasites had smaller but broadly similar regional predecessors whose features contributed directly to the megasites. The clearest case was Trypillia Phase BII megasites such as Nebelivka, where a process of bricolage of plan elements found separately on earlier sites produced an integrated plan (Reference GaydarskaGaydarska 2020: 36–7). But there was an equally obvious heritage of hillforts of varying sizes prior to the construction of Corneşti and Bil’sk, as well as enclosures of contrasting forms prior to Valencina and settlement sites with burial groups adjacent to house clusters in other, earlier Lengyel sites. The principal difference between the megasites and earlier manifestations was a vastly increased scale and intensity of action, which gave a profoundly different experience to visitors and residents alike. However, we can hardly claim a similar socio-political structure for all the megasites, with the antecedents of Iron Age Bil’sk already megasites in their own right and displaying signs of political centralisation far in excess of megasite antecedents in the Neolithic and Copper Age. The Late Bronze Age megaforts of the Carpathian Basin stood in an intermediate position, with a three-level site hierarchy (Reference Molloy, Jovanović and BruyèreMolloy et al. 2020) of greater complexity than the Lengyel or Iberian sites. Yet megasites with broadly similar biographical trajectories developed out of each of these contrasting community groupings and power structures. How did this work?

7.4 Megasite Biographies (Table 7)

The main trends in the site biographies of the five megasites form a polythetic set (Table 7), with the presence of most elements at most megasites the confirmation of a distinct site class (OA/Clarke 1968: 37–8). This is not to claim homogenous social practices across all megasites: the absence of a mortuary zone in Trypillia megasites and Corneşti contrasts strongly with all other congregation places, while there are variable signs of local production using specialised skills (pottery at Trypillia sites, ivory and copper at Valencina, several metals at Bil’sk but no evidence for bronze metallurgy at Corneşti). Nonetheless, the centuries-long duration of all the megasites stimulated the accumulation of place-value, despite the inevitable tensions between strangers and kinsfolk and between the strategic advantages of mortuary and domestic discard, as manifested in the contrastive use of space, noted especially at Valencina.

The long duration of the megasites had another implication – the potential change in site size and use through time, as exemplified by expanding concentric enclosure plans (Perdigões and Corneşti) and the expansion or shrinkage of inner open areas on Trypillia megasites. The development of what were in effect internal antecedent features in the early megasite phases could have over-ridden other landscape features through difference (e.g., the houses of the dead at Alsónyék) or consolidated the local site heritage through emulation (e.g., the Eastern and Western hillforts at Bil’sk). These cumulative changes in morphology reinforced ‘collective megasite identity’ through the conjoint increase in fame of the people and their place (cf. OA/Weiner 1992), as well as the reinforcement of cultural memory enchained to megasite locales. The deposition of the largest, if undated, hoard of spearheads found at Valencina late in its sequence illustrates the impact of long-term fame and vast spatial range on the practices of future generations. The production of cultural memory was particularly striking in the emotional peaks of mortuary performances, whether the killing of cattle for ritual feasting at Alsónyék, the conjoint sensory deprivation and stimulation of entering a Valencina tholos or the participation in burial rites in a monumental barrow at Bil’sk. The cultural memory production through expanded scale of feasting was common to all megasites.

One of the most unexpected findings of this study was the discovery that megasites constituted a cultural storehouse of most, if not all, of the site and monument classes found in the wider landscape. While the temporality of this process is best understood in a preliminary way at Valencina, the cumulative recapitulation of important site classes or elements helped to sustain all megasites over the long-term by making the unknown more familiar and by strengthening the regional network linking smaller sites to the megasites. The puzzling absence of the important antecedent congregational feature at Alsónyék – the Rondel – although present at the nearby Zengővárkony complex can perhaps be explained by the greater emphasis on the dynamics of mortuary congregation at the former.

The other summary statement made at megasites concerned the deposition of the full range of known exotic goods and materials. This can be clearly demonstrated at Alsónyék, Valencina and Bil’sk, while the evidence is lacking at Corneşti and there was a general antipathy to the deposition of exotic polished stone and metal in the worlds of clay comprising Trypillia megasites. It is surely not coincidental that the only place known so far for deposits of both Asian and African elephant ivory was also the largest congregation place known so far in the Iberian Copper Age. The ‘local’ catchment of the megasites for building and other heavy materials varied from 20–100 km, with the bodies of the deceased brought to mortuary congregation places such as Alsónyék or the megaliths of Valencina or Perdigões from perhaps 20–30 km. We propose that the import of more remote materials and things passed through two stages – indirect import through various sites in the ‘local’ catchment, which led to a strengthening of the local networks, followed by direct import by long-distance specialists to the megasite. Rather than posit a link between peaks in exotic networks and the creation of megasites, we would favour a reciprocal attraction, as in the build-up of Greek pottery at Bil’sk.

If the incorporation of summary statements of sites and monument types and exotic exchange networks was a major factor in integrating strangers and kinsfolk at the megasites, how did their personal experiences contribute to the rise, maintenance and decline of European megasites?

7.5 Scalar advantages and disadvantages of megasites: a summary

At the heart of any congregational place lay personal meetings. Megasites made possible different kinds of interactions between people, households, lineages and communities. People could meet others much more regularly, as well as meeting people they had never met before. Thus, relations of trust and friendship would have developed, with deeper relations through time leading not only to inter-marriage but group alliances and opportunities for political fusion. These developments would have had positive feedback effects on exchange networks, with the attraction of ever more exotic finds to specific megasites. It seems inescapable that the trial-and-error of organising hitherto unimagined congregations would have led to successes and failures, with failed congregation sites hard to identify or distinguish from short-term settlement sites. But what makes the megasites studied here remarkable is the length of time over which they continued in use – if anything, marking generations of successful meetings. Given what could (and often did) go wrong on megasites, the long-term success of certain megasites is quite remarkable and requires recognition as the occasional triumph of sociality over increasing social entropy and the autarky and relatively short settlement occupations characteristic of much of European prehistory.

The disadvantages that megasite populations may have encountered have been summarised under the umbrella term ‘scalar stress’, including pollution, noise and illness, friction over decision-making and tensions over global–local identities. Moreover, the instabilities and even failures of long-distance exchange would have diminished one of the principal attractions of a megasite – the availability of a wider range of exotics than anywhere else. Indeed, the possibilities of switching allegiance to a rival congregation centre with perhaps inflated offers of potential new interactions may have damaged the prospects of a megasite with a traditional visitor base, as we suggested could have happened to contribute to the abandonment of Nebelivka in favour of Majdanetske or Taljanki. Nonetheless, we need to identify specific aspects of megasite abandonment rather than make bland assumptions about the inevitability of increased social entropy.

7.6 Conclusions

We believe that we have demonstrated sufficient commonalities between the European prehistoric megasites that we have studied to suggest that they embodied a coherent class of site, in which the best explanation of their size and scale is as particularly large examples of congregation places. Despite the long-term tensions between local relatively independent communities and congregation places, the fact that megasites demonstrated the summary statements of the full range of both sites and monuments and exotic imports showed that the populations in their catchments were well integrated into the central place – if not continuously, then often for long periods of time. It was this repeated, cumulative success in social integration that makes megasites such an important part of European prehistory. If this Element has not only improved our conversation with megasites in European prehistory but also strengthened the case for their future research funding, we can rest content.